Trends and predictors of HIV-1 acquired drug resistance in Minas Gerais, Brazil: 2002–2012

w w w . e l s e v ie r . c o m / l o c a t e / b j i d

The

Brazilian

Journal

of

INFECTIOUS

DISEASES

Original

article

Trends

and

predictors

of

HIV-1

acquired

drug

resistance

in

Minas

Gerais,

Brazil:

2002–2012

Helena

Duani

_,

_Agdemir

_Waleria

_Aleixo

_,

_Unaí

_Tupinambás

a_{UniversidadeFederaldeMinasGerais(UFMG),FaculdadedeMedicina,DepartamentodeClínicaMédica,BeloHorizonte,MG,Brazil} b_{UniversidadeFederaldeMinasGerais(UFMG),FaculdadedeMedicina,Pós-Graduac¸ãoemCiênciasdaSaúde,InfectologiaeMedicina}

Tropical,BeloHorizonte,MG,Brazil

c_{UniversidadeFederaldeMinasGerais(UFMG),LaboratóriodeImunologiaeBiologiaMolecular(DIP-UFMG),BeloHorizonte,MG,Brazil} d_{UniversidadeFederaldeMinasGerais(UFMG),HospitaldasClínicas,BeloHorizonte,MG,Brazil}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received21July2016 Accepted30November2016 Availableonline23December2016

Keywords: HIV Antiretroviral Resistance Genotyping Epidemiology

a

b

s

t

r

a

c

t

Severalstudiesshowthattheprevalenceofmultidrug-resistantHIV-1virusisdecliningover time.Aretrospectivecohortstudywascarriedouttoevaluatethetrendsofdrugresistance inantiretroviraltreatment-exposedindividualsinastateofamiddle-incomecountry,Minas Gerais,southeastregionofBrazil.Weanalyzed2115HIV-1sequencesfrom2002upto2012, from52citiesofMinasGerais.Thegroupswereanalyzedaccordingtothedefinitions:“IAS –3classmutations”,if≥1drugresistancemutationfromIAS2015list(DRM)waspresent ineachclass;“Nofullysusceptibledrugs”astheabsenceofanyfullysusceptibledrugin Stanfordalgorithm;and“GSS≥2,whenamaximumcalculatedGSS(genotypic suscepti-bilityscore)was≥2or≥3,countingonlydrugsavailableinBrazilandUSAatgivencalendar years.TimetrendsofresistancewereanalyzedbyCochran–Armitagetest.Weobserveda decreaseintherateresistancemutationsforPI,NRTI,“IAS–3classmutations”,and“No fullysusceptibledrugs”overthese11years,from69.2%to20.7%,92.3%to90.2%,46.2%to 22.5%,and12.8%to5.7%,respectively(p<0.05).ResistancetoNNRTIincreasedfrom74.4% to81.6%,mainlybecauseofK103Nmutation.TheGSSscore≥2increasedduringtheyears from35.9%to87.3%(p<0.001).WedemonstratethatresistancetoPIandtothethreemain classessimultaneouslyaredeclining,althoughthenumberofpatientsonofantiretroviral therapyhasdoubledinthelasttenyearsinBrazil(125,000in2002to400,000in2014).Broader resistancetestingandtheavailabilityofmoretherapeuticoptionsmighthaveinfluenced thisdecline.TheincreaseinNNRTIresistancecanlimitthisclassasfirstlinetreatmentin Brazilinthefuture.

©2016SociedadeBrasileiradeInfectologia.PublishedbyElsevierEditoraLtda.Thisisan openaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/ by-nc-nd/4.0/).

∗ _{Correspondingauthor.}

E-mailaddress:hduani@gmail.com(H.Duani).

http://dx.doi.org/10.1016/j.bjid.2016.11.009

1413-8670/©2016SociedadeBrasileiradeInfectologia.PublishedbyElsevierEditoraLtda.ThisisanopenaccessarticleundertheCC BY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.0/).

treatment(ART)becameavailable,drasticchangesoccurred in the management of HIV-infected patients. More than 11 million people were on treatment by the end of 2013 withsubsequentimprovementinpatientsqualityoflifeand survival.1_{However,thisincreaseinuseofARTalsoincreased}

acquiredandtransmitteddrugresistance(TDR).The preva-lenceofacquiredHIV-1drugresistance(ADR)inhigh-income countriesfrom Europeand North America,whereART has beenwidelyavailable,isdecliningovertime.2–5_DeLucaetal.

showedthat inWesternEurope countries,three-classdrug resistanceincreasedfrom 0% in1997to apeak of3.0% in 2005,anddeclinedto2%in2008.6_{AnotherstudyfromPortugal}

showedadecreaseintheprevalenceofmultidrugresistance over the last decade, from 6.9% in 2001–2003 to 0.6% in 2009–2012.3_{However,whenanalyzingNNRTIclass,the}

major-ityofthe studiesshowedanincreaseindrugresistance.6–8

There are many studies of ADR from middle-income and resource-limitedcountries;however,longitudinalsystematic dataare scarce.9 _{Consideringthis scenario,theaimofthis}

studywastoanalyzethetrendandidentifythepredictorsof acquiredHIVresistanceinthestateofMinasGerais,southeast regionofBrazil.MinasGeraisisalargestatewith20.6million inhabitantsandmorethan42,000casesofAidsupto2015. Since1996, the Brazilianfederal government provides HIV treatmentfree-of-chargetomorethan400,000patientsanda rationaluseofresourcesisimportanttoensuretheprogram’s sustainability. Moreover, the emergence of viral resistance duringHIV-1infectionmayjeopardizetheBrazilianpolicyfor firstlineandsalvageregimenscurrentlyproposed.Thus,in ordertoassessdrugresistanceprofilesandimprove antiretro-viralstrategiesthesurveillanceofacquireddrugresistanceis substantial.Thisstudycoversawideperiod,fromthemoment thatresistancetestingbecameroutineinclinicalpracticein ourcountryin2002–2012,andcanprovidevaluable informa-tionaboutpredictorsofacquireddrugresistanceinthestate ofMinasGerais.

Material

and

methods

AllHIV-1genotypingtestsofthepublichealthsystemofMinas Geraisstateareperformedinourlaboratory,Laboratóriode ImunologiaeBiologiaMolecularDIPattheSchoolofMedicine ofUFMG.Weconsultedourlaboratorydatabasethatcovers 83clinicslocatedover52citiesinthemainmacroregionsof thestateofMinasGerais.Resistancedatawereretrievedfor 2115(69.4%ofallperformedtests)HIV-1-infectedindividuals (adultsandchildren),whohadundergoneatleastonedrug resistancetestingfrom January2002up toDecember2012. Thegeneticdataresultedfrompopulationsequencingusing anautomatedsequencer(ABIPrism3100,AppliedBiosystems) plusacommerciallyavailableassay(ViroSeqHIV-1 Genotyp-ingSystem,v2.0,Abbott)in2002tothefirsthalfof2008and TrugeneHIV-1genotypingassay(SiemensDiagnostics)from 2008onwards.10_{Foreachsequence,thecorrespondingpatient}

demographics,thelastavailableHIVviralload,andCD4+T-cell

lowlevelofresistance,lowlevelofresistance,intermediate resistance,and highlevelofresistance.Thepredicted drug activity wasscored as1, 0.75,0.5,0.25,and0, respectively. Enfuvirtide and raltegravir scored 0if usedprioror during the genotypingtestand1iftheywere neverused. Maravi-roc,rilpivirineandemtricitabinewerenotconsidered,because theywere notavailableinBrazil untilDecember 2012.The obtainedStanford scorescouldthen becomparedwiththe resultsofacohortstudyinvolvingsevencountriesin West-ernEurope.6_{TheyalsousedtheRegasubtypingtool.Thelocal}

EthicsCommitteeapprovedtheproject.

The mutations analyzed were those included in the 2013 IAS-USA list for nucleoside/tide reverse transcrip-tase inhibitors (NRTI), nonnucleoside reverse transcriptase inhibitors (NNRTI), and major resistance mutations to the protease inhibitors (PIs).13 _{The mutations were analyzed}

accordingtothedefinitions6:“IAS–3classmutations”asat least oneDRM present ineach class;“No fully susceptible drugs”astheabsenceofanyfullysusceptibledrugin Stan-fordalgorithm,and“GSS≥2,whenamaximumcalculated GSS (genotypic susceptibility score)was ≥2, counting only drugsthatwerecommerciallyavailableinBrazilatgiven cal-endaryears.GSScalculationwasdoneconsideringallpossible combinationsof:NRTI±NRTI±NNRTI±PI/r±T20±RAL.For this calculation, at most one agent per drug class was allowedwithinacombination,withtheexceptionofNRTIs, in which unlimitedcombinations were allowed, exceptfor stavudine+zidovudineandemtricitabine+lamivudine.6_This

methodofcalculatingGSSandofassigningweights accord-ing to Stanfordalgorithm were followed theway proposed byDeLucaetal.,so thattheresultsofthisstudy couldbe comparedtotheEuropeanstudy,oneofthebiggestpublished dataset.6_{ThedatawereanalyzedusingSPSSv.15(SPSSInc,}

Chicago,IL).Logisticregressionanalysiswasusedtoassess theindependentassociationofthevariablesacrosstime.

Results

Thebaselinedemographiccharacteristics ofparticipantsin the cohortare showninTable1.The2115sequenceswere obtainedfrom 1,887HIV-1-infectedindividualsintreatment failure.Medianageofpatientsatthetimeofgenotypingwas 40.5 years(p25–p7533.4–47.6).Overall,maletofemale ratio was1.6(1311/804).Almost11%ofthepatientshadmorethan onegenotypingtest performed.MedianHIV-RNAviral load was4.3logcopies/mL(p25–p753.9–4.7)andmedianCD4count was226cells/␮L(p25–p7584–368cells/␮L).Themedian num-berofprevioustreatmentsexperiencedbyenrolledsubjects was3(p25–p752–3).Theuseofmonoordualtherapywas1.9% and 30.7%,respectively,and63.3%patientsusedunboosted PI. Non-B subtypes were detectedin30.8% of the subjects includedinthestudyandthesecondmostprevalentsubtype wasF. Allsequencesbelongedtopatientswithahistoryof exposuretoNRTI,72.3%toNNRTI,51%toboostedPI,and3.5% tonoveldrugclass(entryinhibitorsandintegraseinhibitors).

b r a z j i n f e c t d i s . 2 0 1 7; 2 1(2) :148–154

Table1–Baselinedemographic,immunovirologicalandARThistoryfeaturesofpatientsintherapeuticfailureinMinasGerais,Brazil:2002–2012.

Characteristic Percent(%)b _pb

Allyears 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

Analyzedsequences(N) 2115 39 77 171 183 279 120 291 284 252 175 244

Age(years)c _{40.5(33.4–47.6)}a _{35.6 38.6 38.3 40.4 39.6 36.9 39.6 42.6 42.5 41.3 43.2 <0.001}

Sexmale(N.%) 1311(62.0) 74.4 75.3 69.0 65.0 63.8 60.8 63.2 57.4 56.3 59.4 58.6 <0.001

NumberofARTregimensbefore genotyping

3(2–4)a _{3.0 3.0 2.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 2.0 0.055}

Prioruseofmonotherapywith zidovudine(N.%)

40(1.9) 5.1 7.8 5.8 1.6 4.3 5.0 0.3 0.0 0.0 0.0 0.0 <0.001

Prioruseofdualtherapy(N.%) 650(30.7) 51.3 57.1 42.1 42.6 39.4 24.2 27.5 28.2 21.0 25.1 16.4 <0.001

UseofunboostedPI-basedtherapy(N. %)

1339(63.3) 82.1 84.4 73.7 73.2 71.0 71.7 65.3 60.9 57.5 56.0 37.7 <0.001

UseofboostedPI-basedtherapy(N.%) 1079(51.0) 38.5 36.4 36.3 41.5 39.4 45.0 59.5 59.2 58.7 55.4 60.7 <0.001

PrioruseofNNRT(N.%) 1528(72.3) 56.4 49.4 62.0 65.0 59.9 69.2 81.1 77.1 79.0 78.3 82.8 <0.001

PrioruseofT20(N.%) 48(2.3) 0.0 0.0 0.0 0.0 0.4 0.8 7.9 2.5 4.0 1.7 1.2 <0.001

PrioruseofRAL(N.%) 24(1.1) 0.0 0.0 0.0 0.0 0.0 0.0 2.7 1.1 2.0 1.1 2.5 0.014

PrioruseofETR(N.%) 2(0.1) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.8 0.0 0.0 0.140

BaselineCD4+T-cellcount(cells/␮L)c _{226.0(84–368)}a _{235.0 196.0 220.0 196.0 218.0 259.5 268.0 218.5 210.5 239.0 210.0 0.062}

BaselineHIV-1plasmaviralloadc (log10copies/mL) 4.3(3.9–4.7)a _{4.7 4.5 4.4 4.4 4.3 4.4 4.2 4.2 4.1 4.0 3.9 <0.001} Subtype(N.%) B 1294(61.2) 64.1 70.1 67.8 71.6 62.7 63.3 55.0 62.0 57.9 57.1 55.3 <0.001 F 463(21.9) 17.9 20.8 22.8 19.1 24.4 20.8 24.1 18.3 19.4 24.6 24.2 0.647 a _{Values:median(p25–p75).} b _{Cochrane–Armitagetestfortrends.} c _{Pearsoncorrelationfortrends.}

80 70 60 50 40 30 20 10 0 2002 2003 2004 2005 2006 2007 PI NRTI NNRTI %

IAS 3 class mutations No fully susceptible drug GSS≥2

2008 2009 2010 2011 2012

Fig.1–TrendsofresistancemutationsamongART-experiencedpatientsinMinasGerais,Brazil:2002–2012(p<0.05forall trends).

UseofanyPI,NNRTI,DRV,ETR,T20,andRALinthelast reg-imenwaspresentin59.9%,36.3%,1.1%,0.1%,1.6%,and0.9% ofthe patients, respectively. Over the years, therewas an increaseofpatientage,offemales,useofNNRTI,non-B sub-types,andadecreaseofviralloadatthetimeofgenotyping. CD4countswerestable.

Drug

resistance

mutations

over

time

Fig.1showstheDRMstrendsovertimetodifferentdrugclass: NRTIassociatedDRMshavedecreasedslightlyfrom92.3%to 90.2%.PImutationswerefoundat69.2%ofpatientsin2002 andat34.8%in2012.NNRTImutationsareincreasing,74.4% in2002to81.6%in2012.“IAS–3classmutations”were46.4% anddecreasedto22.5%in2012.

Thethymidineanalogmutations(TAM):M41L,D67N,K70R, L210W,T215YF,andK219EQdroppedfrom56.4%,41%,20.5%, 41%,69.2%,17.2%to27.9%,19.3%,14.8%,15.6%,38.5%,and 16%,respectively.Asignificantdeclinewasobservedforboth type1TAM(M41L,D67N,L210W,andT215Y)andtype2TAM (K70R,T215F, and K219E/Q) profiles, ranging from 74.4% to 48.4%andfrom71.8%to47.1%,respectively.The69-insertion complexand151complexdisappearedalongtheyears.The K65RandM184Vmutationsincreasedfrom2.6%to7.4%and from69.2%to82%,respectively.AmongNNRTIDRMs,a sig-nificant declinefrom 20.5% to 11.9% was found in Y118IC mutation,andanincreasewasfoundtoK103Nmutation,from 43.6%to48%.ForthePImutations,anincreaseoffrequency

wasfoundforI54ML,L76V,andN88S,withapeakintheyears between2008and2009,subsidingsubsequently.ThePI muta-tions L90MandV82AFTLSshowedasignificant decreasein thisperiod.ThedataareshowninTable2.Themostfrequent mutationsinall years forNRTIwere:M184V(77.4%), M41L (43.5%), and T215Y(39.9%). For NNRTI, G190A, K103N, and V108Iwerepresentin10.7%,42.9%,and11.6%ofsequences, respectively.L90M(24.2%),M46I(20.5%)andV82A(21.1%)were themostcommonDRMstoPI.

Thepercentageofsequencesshowing“Nofullysusceptible drugs”declinedfrom12.8%to5.7%.ThecalculatedGSSofthe salvageregimenaccordingtotheavailabledrugsineachyear increasedfrom35.9%to87.3%,asshowninFig.2.

Themultivariatelogisticregressionanalysistoinvestigate thepredictorsofclassresistancedevelopmentontreatmentis showninTable2.IndependentpredictorsofPIresistancewere genotypingyear ofstudy,gender,CD4count, previousZDV monotherapy,ZDV+ddIdualtherapy,unboostedPI,boosted PI,NNRTI,T20,RAL,andanyPIinthelastregimen.Predictors ofNRTIresistanceweregender,age,CD4count,viralloadat failure,dualZDV+ddItherapy,unboostedandboostedPI,and previousNNRTIuse.NNRTIresistancewasfoundin associa-tionwithsomeindependentpredictors,includinggender,CD4 count,previousunboostedPI,andNNRTIand NNRTI+NRTI useinthelastregimen.Femalegenderwasaprotectivefactor toPI,NRTIandIAS–3classmutations,butnottoNNRTI.The usageofunboostedratherthanboostedPIswasthreetimes morelikelytobeassociatedwithevolutiononPIresistance. Thepredictiveabilityofthelogisticregressionmodelsused

b r a z j i n f e c t d i s . 2 0 1 7; 2 1(2) :148–154

Table2–Predictorsofproteaseinhibitor(PI),nucleosidereverse-transcriptaseinhibitor(NRTI),non-nucleosidereverse-transcriptaseinhibitor(NNRTI),IAS3class resistance,nofullysusceptibledrug,andgenotypicsusceptibilityscore≥2intreatmentfailingHIV-1infectedpatients.MinasGerais,Brazil,2002–2012.

Multivariateanalysis:oddsratio(95%confidenceinterval)

Variable(reference) PI NRTI NNRTI IAS–3class

resistance Nofullysusceptible drug GSS≥2 Year 0.88(0.84–0.92) – – 0.92(0.89–0.96) – 1.24(1.19–1.30) Gender(male) 0.55(0.44–0.69) 0.61(0.44–0.83) 1.37(1.08–1.74) 0.79(0.64–0.97) – 1.61(1.3–1.99) Ageinyears(>50) <13 – 0.51(0.23–1.13) – – – 0.53(0.35–0.82) 13–30 – 0.43(0.23–0.79) – – – – 31–50 – – – – – – CD4count/cells/L(>500) <201 0.5(0.36–0.71) 0.48(0.26–0.9) 1.47(1.05–2.06) – – – 201–500 – 0.79(0.42–1.47) 1.23(0.88–1.71) – – – VLatfailure/copies/mL(>100,000) <10,000 – 1.87(1.14–3.09) – – 0.53(0.34–0.84) – 10,000–50,000 – – – – 0.6(0.39–0.91) – 50,000–100,000 – – – – – – Numberofregimens – – – – – 0.73(0.54–0.99) MonoZDV 2.51(1.01–6.27) – – 2.16(1.08–4.29) 2.83(1.22–6.56) – ZDV+ddI 1.32(1.01–1.73) 1.58(1.02–2.46) – 1.38(1.09–1.75) 1.54(1.08–2.2) 0.52(0.41–0.67)

Otherdualtherapy – – – – – 0.56(0.39–0.82)

UnboostedPI 5.63(4.41–7.2) 1.63(1.16–2.29) 0.72(0.55–0.95) 4.18(3.24–5.4) 2.47(1.6–3.83) 0.64(0.49–0.83) BoostedPI 1.75(1.33–2.31) 0.44(0.31–0.63) – 1.58(1.24–2) 4.4(2.72–7.12) 0.63(0.49–0.81) PrioruseofNNRTI 0.52(0.4–0.68) 0.61(0.4–0.93) 5.57(4.39–7.08) 2.59(2.01–3.33) 1.96(1.33–2.9) 0.53(0.41–0.68)

PrioruseofT20 3.92(1.47–10.43) – – 3.6(1.79–7.2) 3.06(1.61–5.8) 0.14(0.064–0.28)

PrioruseofRAL 3.69(1.05–13.01) – – – – –

NRTI+NNRTIinlast regimen

– – 5.99(4.03–8.92) 0.53(0.4–0.7) 0.44(0.26–0.75) –

AnyPIinthelast regimen

60 50 40 30 20 10 0 2001200220032004200520062007200820092010201120122013 110 100 90 80 60 70 50 40 30 20 10 0 2001200220032004 TDF GSS≥2 (Brazil) GSS≥3 (Brazil) GSS≥2 (USA) GSS≥3 (USA) TDF P ercent, % P ercent, % DRV DRV 200520062007200820092010201120122013

Fig.2–Drugresistanceprofileovertime.Percentageof sequenceswithsalvageregimenwithGSS≥2(blueand gray)andGSS≥3(yellowandorange)accordingtothe availabledrugsinBrazilandUSAeachyear(redline),

p<0.001forallGSStrends.DrugsavailableinBrazil/USA eachyear:1991/1987:ZDV;1993/1991:ddI;1996/1995:3TC, SQV;1996/1996:RTV,IDV;1997/1994:D4T;1998/1996:NVP; 1998/1997:NFV;1999/1998:EFV;1999/1997:DLV;2001/1998: ABC;2002/2000:LPV/r;2003/2001:TDF;–/2003:FTC; 2004/2003:ATV;2005/2003:FPV,T20;2008/2006:DRV; 2009/2007:RAL;2009/2005:TPV;2010/2008:ETR;–/2011: RPV;2013/2007:MVC.

showedthatmutationsofresistancetoNNRTIandPIwere bet-terpredictedbytheconsideredvariables,astheyhadgreater sensitivityandspecificity.

Discussion

This study describes the results of the HIV genotyped sequencesofARTfailingpatientsreceivingcareinthestate ofMinasGerais,Brazilduring11years.Thedemographicdata arerepresentativeofHIV-infectedpatientsinthewhole coun-try,withmedianageof40yearsandapredominanceofmale sex.14_{Anincreasinginageofthefailingsubjectswasobserved}

acrossthe years showingthat patientsattendingHIV clin-icsaregettingolderduetolongersurvivalasresultofmore potentandwelltoleratedfirst-lineregimens.Asamatterof factthe incidence ofHIVinfection ishigher amongyoung peopleinBrazil.15_{Ofnote,thedecreaseinviralloadatthe}

timeofgenotypingcouldbeassociatedtoearlierdetection ofthetherapeuticfailure.16_{Themultivariateanalysisshowed}

ofBrazilaremainlymenwhohavesexwithmen(43%).This subsetofpatientsishistoricallymoreexposedtosuboptimal therapythantheoverallpopulation.17,18_{Datafromalarge}

Ital-ianstudyshowsadifferentpattern,withfemalegenderbeing atahigherrisktodevelopresistancemutationstoPI,NRTI, andNNRTIclasses.4

The prevalence ofF subtype is increasing over time in ourstate.TheHIVepidemicinSouthAmericaandBrazilis mainlycausedbysubtypeB.ThesubtypeCpredominatesin the southofthecountry andneighboringcountries, reach-ing 44%insomecities.19–22 _{ThesubtypeFpredominatesin}

Northeastofthecountry.23_{Phylogeneticanalysisisnecessary}

tounderstandthedynamicsoftheFsubtypeincreaseinMinas Gerais.

Wefoundthatinthis cohortofART-exposedgenotyped patientsinamiddle-incomecountry,thePI,NRTI,andIAS–3 classresistanceisdecliningovertime,whereasNNRTI resis-tanceisincreasing.ThesedatawereconfirmedtoPI,IAS–3 classmutations,andGSS≥2bymultivariateanalysis show-inganindependentassociationofcalendaryearwithreduced probabilityofresistance.National dataconfirmthistrend24

and thisisinlinewiththe NationalHIVTreatment Guide-lineschangesduringtheseyears,whichencouragedtheuse ofNRTI+NNRTI(EFV)asthefirsttreatmentoption.14,25 The NNRTI resistanceincreasewasalsoassociatedtothelower geneticbarrierofthisdrugclass.From2002to2005,themost commonNNRTI mutation,K103N,decreased,but startedto increase after 2005 through to 2012. Our state data differ from theNationaldata: whiletheof3classresistancerate inMinasGeraistendstodecrease,thisrateremainedstable countrywide.16

Data from high-income countries from Europe show similar trends to PI, NRTI, and IAS – 3 class resistance mutations.3,6,26,27 _{Somehigh-incomecountriesalsoshowed}

thattherateofNNRTIresistancetendstoincrease.3,6,27,28_For

thePIclass,thisdeclinewasofalmost50%,anditcanbethe reflexofthemorewidespreaduseofboostedPI.

ThereductionofADRatthemomentoftherapeuticfailure canalsobeassociatedtothereductionofTDR,29,30_although

inaGermanstudytheproportionofTDRwasstabledespite asignificantADRdeclineovertime.27_{TheTDRinBrazilhas}

beenstudiedbutnotsystematically,sothereareno longitu-dinaldatatocompareTDRandADR.17,23_{Arecentstudyhas}

shownthattheremaybesomeassociationbetweenresistance mutationsthatappearamongpatientsfailingtherapy(ADR) andTDRinpatientswitharecentinfection,suggestingthat theADRcanalsobeusedtomonitorTDR.16

Themainlimitationofthisstudyisitsretrospectivedesign, basedonthesampleofasinglestateofBrazil.Thenational databanksandmedicalformsusedtoextractbaselinedatadid notcontainallrequestedinformationandmissingdatawere common,thuslimitingthesizeofthesample.Adherencewas animportantvariablethatcouldnotbeevaluatedhere.

In summary, we have demonstrated that antiretroviral resistanceisdeclining inMinasGerais,Brazil,althoughthe ARVavailabilityhasmorethantripledinthelast10yearsin

thecountry(125,000in2002to400,000in2014).Broader resis-tancetestingmighthaveinfluencedthisdecline,aswell as resistancetestingguidedtherapyandtheavailabilityofmore therapeuticoptions.

Conflicts

of

interest

Theauthorsdeclarenoconflictsofinterest.

Acknowledgements

The Projeto G.E.R.A. is team: Ronaldo Batista Melo Junior, LuanaAssis Torres,LuisaCinque de Melo,IsabelaVianello Valle,LigiaIasminePereiradosSantosGualberto,JúliaTeixeira Tupinambás,andMarianaFausterFonsecaAraújodeOliveira.

r

e

f

e

r

e

n

c

e

s

1. WHO.GlobalupdateonthehealthsectorresponsetoHIV; 2014.Availablefromhttp://appswhoint/iris/bitstream/10665/ 128494/1/9789241507585engpdf?ua=12014[accessed 24.12.14].

2. CharestH,Doualla-BellF,CantinR,etal.Asignificant reductioninthefrequencyofHIV-1drugresistanceinQuebec from2001to2011isassociatedwithadecreaseinthe monitoredviralload.PLOSONE.2014;9:e109420.

3. VercauterenJ,TheysK,CarvalhoAP,etal.Thedemiseof multidrug-resistantHIV-1:thenationaltimetrendin Portugal.JAntimicrobChemother.2013;68:911–4.

4. FranzettiM,ViolinM,AntinoriA,etal.Trendsandcorrelates ofHIV-1resistanceamongsubjectsfailinganantiretroviral treatmentoverthe2003–2012decadeinItaly.BMCInfectDis. 2014;14:398.

5. AbrahamAG,LauB,DeeksS,etal.Missingdataonthe estimationoftheprevalenceofaccumulatedhuman immunodeficiencyvirusdrugresistanceinpatientstreated withantiretroviraldrugsinNorthAmerica.AmJEpidemiol. 2011;174:727–35.

6. DeLucaA,DunnD,ZazziM,etal.Decliningprevalenceof HIV-1drugresistanceinantiretroviraltreatment-exposed individualsinWesternEurope.JInfectDis.2013;207:1216–20.

7. vonWylV,YerlyS,BurgisserP,etal.Long-termtrendsofHIV type1drugresistanceprevalenceamongantiretroviral treatment-experiencedpatientsinSwitzerland.ClinInfect Dis.2009;48:979–87.

8. VercauterenJ,WensingAM,vandeVijverDA,etal.

Transmissionofdrug-resistantHIV-1isstabilizinginEurope.J InfectDis.2009;200:1503–8.

9. WHO.SurveillanceofHIVdrugresistanceinadultsreceiving ART.Geneva,Switzerland:HIV/AIDSDo;2014.

10.GrantRM,KuritzkesDR,JohnsonVA,etal.Accuracyofthe TRUGENEHIV-1genotypingkit.JClinMicrobiol.

2003;41:1586–93.

11.AlcantaraLC,CassolS,LibinP,etal.Astandardized frameworkforaccurate,high-throughputgenotypingof recombinantandnon-recombinantviralsequences.Nucleic AcidsRes.2009;37:W634–42.

12.deOliveiraT,DeforcheK,CassolS,etal.Anautomated genotypingsystemforanalysisofHIV-1andothermicrobial sequences.Bioinformatics.2005;21:3797–800.

13.JohnsonVA,CalvezV,GunthardHF,etal.Updateofthedrug resistancemutationsinHIV-1:March2013.TopAntivirMed. 2013;21:6–14.

14.MinistériodaSaúde.In:AidsPNdDe,editor.Protocoloclínico ediretrizesterapêuticasparamanejodainfecc¸ãopelohivem adultos.Brasília,Brasil:MinistériodaSaúde.Secretariade VigilânciaemSaúde;2013.

15.MinistériodaSaúde.In:ViraisDdDAeH,editor.Boletim Epidemiológico–AidseDSTAnoII–no1–atésemana epidemiológica26a.Brasília,DF:MinistériodaSaúde, SecretariadeVigilânciaemSaúde;2013.

16.TilghmanMW,Perez-SantiagoJ,OsorioG,etal.Community HIV-1drugresistanceisassociatedwithtransmitteddrug resistance.HIVMed.2014;15:339–46.

17.TupinambasU,DuaniH,MartinsAV,AleixoAW,GrecoDB. Transmittedhumanimmunodeficiencyvirus-1drug resistanceinacohortofmenwhohavesexwithmeninBelo Horizonte,Brazil–1996–2012.MemInstOswaldoCruz. 2013;108:470–5.

18.GoldsamtLA,ClattsMC,ParkerMM,ColonV,HallackR, MessinaMG.Prevalenceofsexuallyacquiredantiretroviral drugresistanceinacommunitysampleofHIV-positivemen whohavesexwithmeninNewYorkcity.AIDSPatientCare STDs.2011;25:287–93.

19.BrindeiroRM,DiazRS,SabinoEC,etal.BrazilianNetworkfor HIVDrugResistanceSurveillance(HIV-BResNet):asurveyof chronicallyinfectedindividuals.AIDS.2003;17:1063–9.

20.BrennanCA,BritesC,BodelleP,etal.HIV-1strainsidentified inBrazilianblooddonors:significantprevalenceofB/F1 recombinants.AIDSResHumRetroviruses.2007;23:1434–41.

21.MachadoLF,IshakMO,VallinotoAC,etal.Molecular epidemiologyofHIVtype1innorthernBrazil:identification ofsubtypesCandDandtheintroductionofCRF02AGinthe AmazonregionofBrazil.AIDSResHumRetroviruses. 2009;25:961–6.

22.GrafT,PintoAR.TheincreasingprevalenceofHIV-1subtype CinSouthernBrazilanditsdispersionthroughthecontinent. Virology.2013;435:170–8.

23.deMoraesSoaresCM,VergaraTR,BritesC,etal.Prevalenceof transmittedHIV-1antiretroviralresistanceamongpatients initiatingantiretroviraltherapyinBrazil:asurveillancestudy usingdriedbloodspots.JIntAIDSSoc.2014;17:19042.

24.DiazRS,InocencioLA,SucupiraMC,etal.Thevirologicaland immunologicalcharacteristicsoftheHIV-1-infected

populationinBrazil:frominitialdiagnosistoimpactof antiretroviraluse.PLOSONE.2015;10:e0139677.

25.MinistériodaSaúde.In:AidsPNdDe,editor.Recomendac¸ões paraTerapiaAnti-retroviralemAdultosInfectadospeloHIV 2008.7aEdic¸ão.ed.Brasília,Brasil:MinistériodaSaúde. SecretariadeVigilânciaemSaúde;2008.

26.VercauterenJ,DeforcheK,TheysK,etal.Theincidenceof multidrugandfullclassresistanceinHIV-1infectedpatients isdecreasingovertime(2001–2006)inPortugal.Retrovirology. 2008;5:12.

27.SchmidtD,KollanC,FatkenheuerG,etal.Estimatingtrends intheproportionoftransmittedandacquiredHIVdrug resistanceinalongtermobservationalcohortinGermany. PLOSONE.2014;9:e104474.

28.PaquetAC,SolbergOD,NapolitanoLA,etal.Adecadeof HIV-1drugresistanceintheUnitedStates:trendsand characteristicsinalargeprotease/reversetranscriptaseand co-receptortropismdatabasefrom2003to2012.AntivirTher. 2014;19:435–41.

29.YangWL,KouyosR,ScherrerAU,etal.Assessingtheparadox betweentransmittedandacquiredHIVType1drugresistance mutationsintheSwissHIVcohortstudyfrom1998to2012.J InfectDis.2015.

30.PhamQD,WilsonDP,LawMG,KelleherAD,ZhangL.Global burdenoftransmittedHIVdrugresistanceandHIV-exposure categories:asystematicreviewandmeta-analysis.AIDS. 2014;28:2751–62.