Can HIV reverse transcriptase activity assay be a low-cost alternative for viral load monitoring in resource-limited settings?

Can HIV e e se ansc ip ase ac i i y assay be a low-cos al e na i e o i al load moni o ing in esou ce-limi ed se ings? Soham Gup a, 1 Riya Palchaudhu i, 1 Ujjwal Neogi, 2 Hi esa e S ini asa, 1 Pe Asho n, 3 Ayesha De Cos a, 4 Clas Källande , 5 Ani a She 6 To ci e: Gup a S, Palchaudhu i R, Neogi U, e al. Can HIV e e se ansc ip ase ac i i y assay be a low-cos al e na i e o i al load moni o ing in esou ce- limi ed se ings?. BMJ Open 2016;6:e008795. doi:10.1136/bmjopen-2015- 008795 ▸P epublica ion his o y o his pape is a ailable online. To iew hese iles please isi he jou nal online (h p://dx.doi.o g/10.1136/ bmjopen-2015-008795). SG and UN a e co- co esponding au ho s. Recei ed 28 May 2015 Re ised 16 No embe 2015 Accep ed 27 No embe 2015 Fo numbe ed a ilia ions see end o a icle. Co espondence o Soham Gup a; [email p o ec ed] and Ujjwal Neogi; ujjwal.neogi@ ki.se ABSTRACT Objec i e: To e alua e he pe o mance and cos o an HIV e e se ansc ip ase-enzyme ac i i y (HIV-RT) assay in compa ison o an HIV-1 RNA assay o ou ine i al load moni o ing in esou ce limi ed se ings. Design: A coho -based longi udinal s udy. Se ing: Two an i e o i al he apy (ART) cen es in Ka na aka s a e, Sou h India, p o iding ea men unde he Indian AIDS con ol p og amme. Pa icipan s: A coho o 327 HIV-1-in ec ed Indian adul pa ien s ini ia ing i s -line ART. Ou come measu es: Pe o mance and cos o an HIV-RT assay (ExaVi Load V3) in compa ison o a gold s anda d HIV-1 RNA assay (Abbo m2000 ) in a coho o 327 Indian pa ien s be o e (WK00) and 4 weeks (WK04) a e ini ia ion o i s -line he apy. Resul s: Plasma i al load was de e mined by an HIV- 1 RNA assay and an HIV-RT assay in 629 samples (302 pai ed samples and 25 single ime poin samples a WK00) ob ained om 327 pa ien s. O e all, a s ong co ela ion o =0.96 was obse ed, wi h good co ela ion a WK00 ( =0.84) and a WK04 ( =0.77). Bland-Al man analysis o all samples showed a good le el o ag eemen wi h a mean di e ence (bias) o 0.22 log 10 copies/mL. The pe o mance o ExaVi Load V3 was no nega i ely a ec ed by a ne i apine/e a i enz based an i e o i al egimen. The pe es cos o measu ing plasma i al load by he Abbo m2000 and ExaVi Load V3 assays in a basic lab se ing was $36.4 and $16.8, espec i ely. Conclusions: The s ong co ela ion be ween he HIV- RT and HIV-1 RNA assays sugges s ha he HIV-RT assay can be an a o dable al e na i e op ion o moni o ing pa ien s on an i e o i al he apy in esou ce-limi ed se ings. T ial egis a ion numbe : ISRCTN79261738. INTRODUCTION The p incipal aim o an i e o i al he apy (ART) is du able supp ession o eplica ing plasma i us o unde ec able le els, he eby delaying disease p og ession and p olonging su i al. 12 Expanding access o ART in esou ce-limi ed se ings along wi h close moni o ing is needed o success ul ea - men ou comes. In high income se ings, his is achie ed by pe o ming quan i a i e i al load moni o ing e e y 3–6 mon hs, 3 as i al load moni o ing de ec s ea ly ea men ailu e. Howe e , in esou ce-limi ed se ings, he apeu ic ou come is e alua ed ei he on he basis o a CD4 T cell coun o clinical findings, 4 nei he o which accu a ely p e- dic s i al supp ession. 5 Ea ly de ec ion o i al ailu e by moni o ing he i al load also p o ides he oppo uni y o in ensi y adhe - ence counselling o imp o e adhe ence o ART, po en ially leading o esupp ession o i al load be o e he e olu ion o d ug- esis an i us can ake place. 6 The cu en ly used i al load assays a e based on he amplifica ion o HIV-1 i ion S eng hs and limi a ions o his s udy ▪To he bes o ou knowledge, he cu en s udy is he mos ho oughly e alua ed s udy o ExaVi Load V3 HIV e e se ansc ip ase-enzyme ac i - i y (HIV-RT assay) om India o da e. ▪This s udy was pe o med in a la ge numbe o pa ien s in a longi udinal manne looking in o he e ec s o he non-nucleoside e e se an- sc ip ase inhibi o s-based he apy and d ug esis ance mu a ions on he pe o mance o he HIV RT-enzyme ac i i y assay. ▪Though he assay has been alida ed be o e in o he non-C sub ype se ings, in his s udy he HIV-RT assay was alida ed on a la ge scale in a sub ype C p edominan se ing and sub ype C is he mos p edominan sub ype globally, mo e so a ec ing he esou ce-limi ed se ings. ▪This s udy was limi ed o pa ien s hailing om Sou he n India and a pe o mance e alua ion and cos -e ec i eness o he HIV-RT assay needs o be accessed on a na ional le el. Gup a S, e al.BMJ Open 2016;6:e008795. doi:10.1136/bmjopen-2015-008795 1 Open Access Resea ch g oup.bmj.com on Ap il 20, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om RNA, which is conside ed imp ac ical o wide-scale use in esou ce-limi ed se ings, as i equi es in as uc u e, acili ies o molecula diagnos ics, expensi e equipmen and skilled echnicians, which a e o en una ailable. 7 Simple , less expensi e i al load assays would be e y use ul in he esou ce-limi ed en i onmen s ha a e mos impac ed by his epidemic. An al e na i e o measu e he HIV-1 RNA is o measu e he ac i i y o he i al e e se ansc ip ase (RT) enzyme. The ExaVi Load assay (Ca idi, AB, Sweden), a low-cos and echnically less demanding assay using an ELISA-based me hod o measu e RT enzyme ac i i y, has shown p omising esul s. 8–19 Al hough p e ious e sions o he es ha e been e alu- a ed agains se e al PCR-based HIV es s ha measu e i al RNA, he e a e ew compa a i e s udies be ween ExaVi Load V3 and such molecula eal- ime assays. 14 16 18 The mos ecen V3 o he ExaVi Load assay has an enhanced sensi i i y (lowe de ec ion limi 200 copies/mL) and educed u na ound ime com- pa ed o V2. 14 18 HIV-1 sub ype C (HIV-1C) is he dominan s ain in mos low and middle income coun ies like India, Sou h A ica and E hiopia, 20 and he need o a simple low cos i al load moni o ing ool is a p io i y in hese se - ings. Da a a ailable om coun ies domina ed by he HIV-1C epidemic a e limi ed. Whe e a ailable, e alua- ions ha e been mainly pe o med wi h ea lie e sion assays (ExaVi Load V1 and V2), which had lowe de ec- ion limi s o 400 copies/mL. 7112122 S udies e alua ing Exa i Load V3 ha e been mainly pe o med in non-C domina ed coun ies. 14 16 18 19 Thus, we aimed o e alua e he pe o mance o he HIV RT-enzyme ac i i y (HIV RT) assay (ExaVi Load V3) in compa ison o a ‘gold s anda d’HIV RNA load (Abbo m2000 eal- ime PCR) assay in a coho o pa ien s be o e and a e ini ia ion o fi s -line ART in Indian se ings. We also compa ed i al load measu e- men s om bo h assays in a subse o pa ien s wi h d ug- esis an mu a ions a baseline. Fu he , we s udied he di e ence in cos s o he wo i al load assays in he con ex o ou labo a o y se ing om a p o ide pe spec i e. MATERIALS AND METHODS S udy pa icipan s and samples Be ween Ap il 2010 and Sep embe 2011, EDTA plasma samples we e collec ed om HIV-1-in ec ed adul pa ien s a ending he In ec ious Disease Clinic, S . John’s Medical College and Hospi al, Bangalo e (main si e) and ART cen e, K ishna Rajend a Hospi al, Myso e (pe iphe al si e) en olled in he HIVIND ando- mised con olled ial (T ial egis a ion: ISRCTN79261738). 23 All he pa ien s included in he s udy ini ia ed ART wi h e e se ansc ip ase inhibi o (RTI) d ugs, ha is, wo nucleoside e e se ansc ip ase inhibi o s (NRTI), zido udine (AZT) o s a udine (d4T) wi h lami udine (3TC) + one non-nucleoside e e se ansc ip ase inhibi o (NNRTI), ei he ne i apine (NVP) o e a i enz (EFV) as pe he s anda d na ional AIDS p og amme guidelines. 24 Plasma samples we e collec ed a wo ime poin s; (1) p io o ART ini ia ion (no longe han 3 mon hs p io ) and (2) 4 weeks a e ART ini ia ion. The plasma samples we e sepa a ed wi hin 6 h o EDTA whole blood collec ion, aliquo ed and s o ed a −80°C in he main si e. Plasma samples aliquo ed in he pe iphe al si e we e s o ed a −20°C and anspo ed o he main si e on d y ice (e e y 2 weeks) and hen s o ed a −80°C p io o es ing. E hical s a emen E hical app o als o he conduc o he ial we e ob ained om he Ins i u ional E hical Re iew Boa d o S John’s Medical College Hospi al, Bangalo e (IERB 1/ 369/08-92/2008) and K ishna Rajend a Hospi al, Myso e (NO/PS/173/2010). All pa ien s pa icipa ing in he HIVIND s udy ha e gi en hei w i en consen . Plasma HIV-1 RNA assay Plasma HIV-1 RNA load was measu ed in he pa ien coho using he Abbo Real-Time PCR, m2000 sys em wi h a manual RNA ex ac ion p ocedu e on an m2000sp sample p epa a ion sys em as pe he manu ac- u e ’s ins uc ions. The assay was pe o med using an ini ial olume o 0.2 mL plasma, which p o ides limi s o quan ifica ion be ween 150 copies/mL (lowe limi o de ec ion) and 10 000 000 copies/mL (uppe limi o de ec ion). In e e y un, a nega i e con ol, a low posi- i e con ol and a high posi i e con ol supplied in he Abbo Real ime HIV-1 con ol ki we e included. This measu e using Real-Time PCR was conside ed as he gold s anda d. This p o ocol was alida ed by an ex e nal quali y con ol p og amme by he Quali y Con ol o Molecula Diagnos ics, Glasgow, Sco land (QCMD, h p://www.qcmd.go ) on a 2010 panel (Consis ing o ou HIV-1B samples, wo HIV-1C samples, one HIV-1A/ G sample and one HIV-1 nega i e sample) and ob ained a highly sa is ac o y sco e. Plasma HIV RT enzyme ac i i y assay The i al RT enzyme ac i i y was quan ified using Ca idi ExaVi Load V3 as pe he ins uc ions o he manu ac- u e . 14 In an ELISA-based o ma , he RT ac i i y o he RT enzyme eco e ed om 1 mL o pa ien plasma was de e mined and compa ed o a s anda d cu e based on known amoun s o an HIV 1 ecombinan RT cons uc ed om he BH10 isola e. The RT ac i i y in he sample was exp essed as em og am HIV-1 RT ac i i y/mL ( g/mL) plasma ha was ansla ed in o RNA copies/mL equi alen by he ExaVi Load analyse so wa e. The lowe de ec ion limi o he assay was ≥1 g/mL ha is, ≥200 copies/mL. In e e y un, a single posi i e con ol and a nega i e con ol we e used o moni o he pe o mance o he RT ex ac ion p ocess 2Gup a S, e al.BMJ Open 2016;6:e008795. doi:10.1136/bmjopen-2015-008795 Open Access g oup.bmj.com on Ap il 20, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om and o access be ween un a ia ions. The posi i e con ol was p epa ed by pooling he EDTA plasma om a high HIV-1 i al load sample and HIV-1 nega i e plasma; 1.2 mL was aliquo ed and s o ed in a −80°C eeze . The labo a o y pe sonnel unning he ExaVi Load V3 assay we e blind o he plasma HIV-1 RNA alues. HIV-1 sub yping and baseline d ug esis ance Geno ypic esis ance es ing (GRT) was pe o med on he baseline plasma samples. B iefly, he e e se an- sc ip ase (RT) egion o he HIV-1 pol gene was ampli- fied and sequenced using he p ime s desc ibed by us p e iously. 25 HIV-1 sub yping was de e mined on he basis o he pol gene as well as he en gene (whe e e sequence da a we e a ailable). 26–28 using he maximum likelihood (ML) phylogene ic ee based on e e ence sequences downloaded om he Los Alamos Da abase (h p://www.hi .lanl.go ). P ima y d ug esis ance ana- lysis was e alua ed using he WHO lis o mu a ions om 2009 (WHO_SDRM 2009). 29 Compa a i e cos analysis We did an analysis compa ing he cos s be ween Abbo m2000 and ExaVi Load V3 om a p o ide (labo a- o y se ice p o ide ) pe spec i e. We used cos s om ou lab o his pu pose. Cos s conside ed included annui ised capi al cos s o he wo di e en ins umen s including ope a o -supplied ins umen s. These a e ins umen s which a e necessa y in case o a new labo a- o y. In case o Abbo , he ope a o -supplied ins u- men s included single-channel mic opipe es, wo d y ba hs and a o ex. Fo ExaVi Load V3, hey included mic opipe es (bo h single-channel and mul ichannel), an ELISA pla e eade , an incuba o , a ocke and a o ex. Cos s o s a -up ki s, human esou ce cos s (includ- ing ime o aining), annual main enance, eagen s and o he consumables we e also conside ed. We assumed he wo king li e o Abbo m2000 and ExaVi Load V3 o be 5 yea s, and a discoun a e o 5% was applied. Assump ions Numbe o pa ien s: The e a e a p esen 1500 ART-expe ienced pa ien s in ou ART cen e. Assuming ha i al load moni o ing o hese pa ien s will equi e o be pe o med e e y 6 mon hs, he e will be 3000 samples a yea . Main enance cos s: In he case o Abbo m2000 he cos s o se icing, main enance o ins umen and cali- b a ion o lase head. In he case o Abbo m2000 he annual main enance cos included he cos s o se - icing, main enance o ins umen and callib a ion o lase heads. Human esou ce skills and aining equi emen s: F om ou expe ience, 1 mon h was equi ed o ain a echnician on Abbo m2000 . One week was equi ed o aining on ExaVi load V3. We also conside ed ha he echnician handling Abbo m2000 would equi e o be mo e senio and expe ienced (sala y $300 pe mon h) compa ed o he echnician wo king wi h ExaVi Load V3 (sala y $200 pe mon h). Time o each me hod o es ing: Fo Abbo m2000 , a ba ch o 24 eac ions which comp ise 21 samples and 3 con ols will in ol e a o al ime o 8 h ( om he beginning o RNA ex ac ion up un il ob aining esul s), o which 5 h in ol e he echnician’s ime. Fo Exa i Load V3, he assay is pe o med in ba ches o 30 samples. Though he u na ound ime is 48 h, i in ol es 5 h o ac ual hands-on ime pe ba ch. Cos s o ime o he echnicians we e calcula ed agains he sala ies men- ioned abo e. In he cos compa ison analysis, we did no conside cos s associa ed wi h he sample collec ion, s o age and anspo a ion as hese a e common o bo h es s. S a is ical analysis All s a is ical analyses we e pe o med a e he HIV RT and HIV-1 RNA le el alues we e log 10 ans o med. Fo analysis, he lowe limi o de ec ion o he HIV RT assay (≤200 copies/mL) was conside ed; samples showing <200 copies/mL by any o he assays we e assigned a alue o 199 copies/mL. Wi h Exa i Load V3, we achie ed a a ying uppe de ec ion limi anging om >360 000 o >770 000 in di e en uns. Thus, samples wi h a i al load o >360 000 ( he lowes ange o he uppe de ec ion limi ob ained o ExaVi Load V3) by any o he assays we e assigned a i al load o 360 000 copies/mL. The diagnos ic ag eemen be ween he HIV RT assay and he HIV-1 RNA assay a di e en i al load cu -o s was de e mined om he κs a is ic. Pea son’s co ela ion coe ficien ( ) was calcula ed o s udy he co ela ion be ween log 10 HIV RT ac i i y (copies/mL equi alen s) and log 10 HIV RNA (copies/mL). Howe e , since his coe ficien does no ake in o accoun he pos- sibili y ha one measu e may di e consis en ly om he o he , we u he assessed he le el o ag eemen using pai wise Bland-Al man plo s. This plo compa es he measu es be ween he wo es s by plo ing he di e - ence in he wo VL measu es agains he a e age o he wo measu es. RESULTS Sample cha ac e is ics Plasma i al load was ob ained om 629 samples col- lec ed om 327 HIV-1-in ec ed adul pa ien s, o which 302 we e pai ed (be o e ART and 4 weeks a e ART ini- ia ion). HIV-1C was he p edominan sub ype obse ed in 98.1% (313/319) o he pa ien s. Six ou o 319 geno- yped pa ien s (1.9%) showed he p esence o non-C sub ype s ains, namely BC ecombinan (1), BD ecom- binan (1), A1C ecombinan (2) and HIV-1A1 (2). Ele en pa ien s (3.4%) showed he p esence o single p ima y d ug- esis an mu a ions, wi h six samples Gup a S, e al.BMJ Open 2016;6:e008795. doi:10.1136/bmjopen-2015-008795 3 Open Access g oup.bmj.com on Ap il 20, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om ha bou ing NRTI-associa ed d ug- esis an mu a ions (DRMs) and fi e samples wi h NNRTI-associa ed mu a- ions ( able 1). Compa ison be ween HIV RT ac i i y and HIV-1 RNA load assay The e we e 54 samples (8.5%) ha we e quan ifiable by he HIV-1 RNA assay bu we e below he de ec ion limi o he HIV RT assay ( able 2). A a lowe limi o quan i- fica ion o 200 copies/mL, 90.7% o he samples showed a quan ifiable i us by he HIV RT assay. The pe cen age o he samples wi h a quan ifiable i al load by he HIV RT assay inc eased wi h highe i al load cu -o s by he HIV-1 RNA assay as shown in able 2. O e all, he e was accep able ag eemen obse ed be ween he HIV RT and HIV-1 RNA assays, wi h excellen ag eemen obse ed a highe alues o plasma i al load ≥3.0 log 10- copies/mL (κ=0.76). O all he samples, 81.7% (514/ 629) had i al load alues by he HIV RT assay, which di - e ed by<0.5 log 10 uni s om he HIV-1 RNA alues, while 99.2% (624/629) o he samples di e ed by <1.0 log 10 uni s. A s ong posi i e co ela ion was obse ed be ween he plasma i al load alues by he HIV RT and HIV-1 RNA assays ( =0.96) in all he samples. A good co ela ion was no ed in ART-naï e samples ( =0.84) as well as in samples a week 04 o ART ( =0.77; figu e 1). Bland-Al man plo s o all he samples showed good le els o ag eemen wi h a mean di e ence (bias) o 0.22 log 10 copies/mL, wi h accep able limi s o ag ee- men (−0.45 and +0.89 log 10 copies/mL). A good le el o ag eemen was also obse ed sepa a ely a baseline (mean di e ence bias o 0.25; ange o accep able limi o ag eemen : −0.39 and +0.89 log 10 copies/mL) and a WK04 (mean di e ence bias o 0.19; ange o accep able limi o ag eemen : −0.52 and +0.89 log 10 copies/mL; figu e 2). In luence o cu en ART and d ug- esis an mu a ions on RT-enzyme ac i i y Table 1 shows ha he mean log 10 di e ence be ween he HIV RT and HIV-1 RNA assays bo h be o e and a e ini ia ion o ART was no significan ly di e en (<0.25 log 10 copies/mL) bu was well wi hin he clinically accep ed limi o 0.5 log 10 copies/mL. Thus, he pe - o mance o he HIV RT assay is no a ec ed by he p es- ence o he NNRTI (ne i apine/e a i enz)-based ART egimen. Also, he p esence o ei he NRTI-associa ed DRMs (n=6; M41L: 1, D67N: 1, T69D: 1, M184I: 1 and T215S: 2) and NNRTI-associa ed DRMs (n=5; Y181C: 1, K101E: 1 and K103N: 2) showed an accep able change Table 1 Compa ison o HIV-1 plasma VL le els measu ed by he HIV RT and HIV-1 RNA assays by ART s a us, HIV-1 sub ypes and RT-d ug esis an mu a ions Sample ype Numbe o samples Mean i al load±SD in log 10 copies/mL Mean log 10 i al load di e ence±SD in log 10 copies/mLExaVi Load V3 Abbo m2000 All samples 629 3.98±1.3 4.19±1.3 0.22±0.3 ART S a us Naï e (baseline a WK00) 327 5.07±0.6 5.33±0.5 0.25±0.3 Expe ienced (WK04) 302 2.79±0.5 2.97±0.6 0.19±0.4 Sub ype a WK00 (n=319)* C 313 5.08±0.6 5.33±0.4 0.25±0.3 Non-C 6 5.05±0.5 5.3±0.4 0.25±0.3 DRMs a WK00 (n=319)* Wild ype (no DRM) 308 5.07±0.6 5.32±0.5 0.25±0.3 NRTI mu a ions 6 5.41±0.3 5.55±0.01 0.15±0.3 NNRTI mu a ions 5 5.26±0.2 5.62±0.2 0.36+0.2 *Geno yping pe o med only in baseline samples. ART, an i e o i al he apy; DRM, d ug- esis an mu a ions, NNRTI, non-nucleoside e e se ansc ip ase inhibi o s; NRTI, nucleoside e e se ansc ip ase inhibi o s; RT, e e se ansc ip ase; VL, i al load. Table 2 Ag eemen be ween he HIV RT assay and HIV-1 RNA assay a di e en PVL le els PVL by HIV-1 RNA In copies/mL (log 10 copies/mL) Ag eemen κValue Numbe o samples de ec ed by Abbo m2000 Pe cen age o Samples de ec ed by ExaVi Load V3 ≥200 (2.3) 89.1 0.46 580 90.7 ≥400 (2.6) 88.1 0.57 550 93.5 ≥1000 (3.0) 89.7 0.76 458 97.6 ≥5000 (3.7) 94.4 0.89 344 99.7 ≥10 000 (4.0) 96.8 0.94 324 100 PVL, plasma i al load; RT, e e se ansc ip ase. 4Gup a S, e al.BMJ Open 2016;6:e008795. doi:10.1136/bmjopen-2015-008795 Open Access g oup.bmj.com on Ap il 20, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om (<0.4 log 10 copies/mL) in mean log 10 di e ence om he co esponding alue among wild ypes. Al hough he samples wi h mu a ions a e small, i indi- ca es ha he p esence o NRTI and NNRTI DRMs did no nega i ely impac he es pe o mance. Cos compa ison o he assays The labo a o y cos o i al load moni o ing o HIV-in ec ed pa ien s analysed in ou coho by bo h Abbo m2000 (HIV-1 RNA) and ExaVi Load V3 (HIV RT) is shown in able 3. The pe es cos o he plasma i al load measu ed by Abbo m2000 and ExaVi Load V3 was $36.4 and $16.8, espec i ely. Thus, by using ExaVi Load V3, $19.6 pe es can be sa ed. In a labo a o y wi h a p e-exis ing basic se -up o ELISA-based assays, ExaVi Load V3 can be pe o med a $16.1, sa ing $20.2 pe es . Mos o he expense sa ed by using ExaVi Load V3 was due o (1) lowe capi al cos s (ins umen s $37 750 agains $2000) (2) he lowe cos o he assay eagen s ($15/ es agains $31/ es ; able 3). Figu e 1 Co ela ion be ween he Abbo m2000 and ExaVi Load V3 assays o (A) all 629 samples showing =0.96 (B) 327 baseline (WK00) samples showing =0.84 and (C) 302 4-weeks pos -ART (WK04) samples showing =0.77. ART, an i e o i al he apy. Figu e 2 Bland-Al man plo wi h 95% CI o limi s o ag eemen be ween HIV-1 i al loads measu ed wi h he Abbo Real-Time m2000 assay and he ExaVi Load V3 assay o (A) all 629 samples showing a mean bias o 0.22 wi h 95% limi s o ag eemen anging om −0.45 o 0.89 (B) 327 baseline (WK00) samples showing a mean bias o 0.25 wi h 95% limi s o ag eemen anging om −0.39 o 0.89 (C) 302 4 weeks pos -ART (WK04) samples showing a mean bias o 0.19 wi h 95% limi s o ag eemen anging om −0.52 o 0.89. ART, an i e o i al he apy. Gup a S, e al.BMJ Open 2016;6:e008795. doi:10.1136/bmjopen-2015-008795 5 Open Access g oup.bmj.com on Ap il 20, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om DISCUSSION A good co ela ion be ween he HIV RT and HIV-1 RNA assays was obse ed in he cu en HIV-1C p edominan se ing in India. The ag eemen be ween he es s was no significan ly a ec ed by he NRTI/NNRTI-based an i e o i al egimen used. Ea lie s udies pe o med on panels o di e en sub ypes and ecombinan s ha e sugges ed ha he HIV RT assay de ec s all HIV-1 and HIV-2 sub ypes wi h simila e ficiency. 14 15 30 This assay, ExaVi Load V3, can he e o e be an a ac i e op ion o i al load moni o ing in Indian se ings. This s udy compa ed he ExaVi Load V3 assay wi h he Abbo m2000 HIV-1 RNA assay and obse ed an excel- len co ela ion ( =0.96). An ea lie s udy om London compa ing he same es s obse ed a simila co ela ion ( =0.94). 14 S ong co ela ions be ween he ExaVi Load V3 and Roche HIV-1 RNA-based assays ha e also been obse ed by wo o he s udies, by G eeng ass e al 18 om Aus alia (Roche Cobas Amplico ; =0.85) and Huang e al 16 om China (Roche Cobas TaqMan 48; =0.95). Nei he o hese s udies was pe o med in HIV-1C domi- na ed se ings. The HIV RT assay showed a good ag ee- men wi h he HIV-1 RNA assay a he clinically impo an i al load h eshold o 1000 copies/mL, which is used by he WHO o define i al ailu e o fi s -line he apy and is also mos o en used as he cu -o o d ug esis ance geno yping. 31 The pe o mance o he Exa i Load below 1000 copies/mL is mode a e. In gene al, we obse ed an unde es ima ion o i al load o 0.22 log 10 RNA copies by he HIV RT assay, which is simila o wha has been obse ed in o he s udies. 14 16 32 These wo su oga e assays use e y di e - en me hods o quan i ying he plasma i al load. The HIV-1 RNA assays quan i y he amoun o i al RNA i e- spec i e o RNA unc ionali y, while he HIV RT assay quan ifies he amoun o ac i e RT enzyme. The calib a ion cons an used o ansla e RT ac i i y in o RNA copies was es ima ed om a s udy o an Aus alian coho 18 and is no comple ely accu a e o all combina- ions o HIV RNA assays and coho s wi h a ying sub ype composi ions. The a ia ion obse ed is, howe e , well wi hin he accep able limi o <0.5 log 10 copies. NNRTI d ugs bind o he RT enzyme, inhibi i s ac i - i y and p e en i al eplica ion. Se e al a icles ha e dis- cussed he possibili y ha enzyma ically inac i e RT d ug complexes could esul in unde quan ifica ion o RT in ela ion o RNA. 29 30 33 These s udies we e, howe e , c oss-sec ional and ne e ound any e idence o educed RT ac i i y du ing NNRTI he apy. In con as o p e ious coho s, he longi udinal sampling in ou s udy p o ides op imal ma e ial o e alua ion o he e ec s o NNRTI-based d ug egimens. When compa - ing HIV i al load da a om he same pa ien coho be o e and a e onse o ART, we ound a mean log 10 di e ence be ween ExaVi Load V3 and Abbo m2000 o 0.25 o naï e pa ien s and 0.19 o expe ienced pa ien s ( able 1). Thus, he di e ence be ween he es s did no inc ease a e onse o he apy. This suppo s e i- dence ha he cu en NNRTI con aining he apy does no ad e sely influence he eco e y o RT enzyme ac i i y. On a small numbe o samples, we assessed i he p es- ence o d ug- esis an mu a ions dec eased he RT fi ness so as o influence he pe o mance o he HIV RT assay. We had six samples wi h single NRTI mu a ions and fi e samples wi h single NNRTI mu a ions and obse ed no e idence ha hei p esence caused any sig- nifican di e ence in he associa ion be ween RT enzyme ac i i y and HIV-1 RNA load. These esul s we e no unexpec ed and suppo e idence om p e ious s udies by Nap a nik e al and Rooijen e al indica ing ha he p esence o NRTI o NNRTI mu a ions do no a ec he ela ionship be ween RT enzyme ac i i y and HIV-1 RNA load. 34 35 Resis ance o NRTIs is media ed by a p imi i e DNA edi ing unc ion ha is in oduced in o he HIV RT by ce ain mu a ions. An ene gy dependen base excision eac ion emo es he las base in he g owing DNA chain. This equi es an ene gy dona o , usually ATP o GTP, and migh dec ease RT eac ion el- oci y. This happens eadily in i o, bu he eac ion con- di ions in he cu en RT assay do no suppo his eac ion. 36 To he bes o ou knowledge, his is he mos ho - oughly e alua ed s udy o ExaVi Load V3 om India o da e. Thus a , he e ha e been h ee compa a i e s udies om India ha ha e been epo ed om he s a es o Andh a P adesh (Anan pu ), 37 Tamil Nadu (Chennai) 7 and New Delhi. 32 Iqbal e al 7 om Chennai c oss-sec ionally e alua ed he ExaVi Load assay V1 and Roche Amplico Moni o assay. They ound a good ag eemen be ween he wo es s and a significan in e se co ela ion be ween ExaVi Load and CD4+ T-cell coun . Al a ez-U ia e al 37 om Anan pu Table 3 Cos compa ison be ween he HIV RT assay and HIV-1 RNA assay o a labo a o y doing 3000 es s/yea Cos i ems ($) Abbo m2000 Cos ($) ExaVi Load V3 Cos ($) Annui ised cos s o capi al ins umen s 8719 462 Annui ised cos o ope a o supplied ins umen s 293 1540 Annual main enance cos 1126 666 Cos s o ki s pe yea 93 000 45 000 Consumables pe yea 4616 1846 T aining ime o lab s a o un he es 300 50 Sala y cos s (3000 es s pe yea ) 1200 1000 To al ($) 109 634 50 534 Cos / es ($) 36.4 16.8 US$1=60 INR. RT, e e se ansc ip ase. 6Gup a S, e al.BMJ Open 2016;6:e008795. doi:10.1136/bmjopen-2015-008795 Open Access g oup.bmj.com on Ap il 20, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om compa ed he accu acy o ExaVi V3 wi h Roche Cobas TaqMan HIV-1 es and Roche Amplico HIV-1 DNA assay o ea ly in an diagnosis. ExaVi pe o med well showing 100% sensi i i y and 99% specifici y, bu no quan i a i e co ela ions we e e alua ed. A mo e ecen s udy by Kokkayil e al 32 compa ed ExaVi Load V3 wi h Roche Cobas TaqMan among 75 ART naï e pa ien s and epo ed no s a is ically significan di e ence be ween he wo assays. The e a e a ew d awbacks o he HIV RT assay. The long u na ound ime o 48 h makes i appea labou in ensi e, hough he ac ual hands-on ime is app oxi- ma ely 5 h. The p olonged incuba ion ime is c i ical o achie e assay sensi i i y. Fo s anda d pe o mance, he assay equi es 1 mL o plasma, which is high in compa i- son o he equi emen o HIV-1 RNA assays, hus limi - ing i s possible use ulness in paedia ic popula ions. Howe e , a ecen s udy by G eeng ass e al 19 obse ed ha sample olumes down o 0.25 mL wi h VL >800 copies/mL can be u ilised o paedia ic moni o ing. The ExaVi Load assay does no p o ide a s anda d posi- i e and nega i e con ol, hus equi ing he lab o p epa e i s own con ols, which may comp omise he quali y assu ance o he assay. Addi ionally, we no ed ha he quali y o he wa e used o washing is impo an as impu i ies and bac e ial con aminan s p esen in wa e may con ain polyme ases which can c ea e backg ound noise and inc ease he le el o he lowe de ec ion limi . In spi e o hese ca ea s, he HIV RT assay has ad an- ages o e he HIV-1 RNA assays in esou ce-limi ed se - ings because i is an ELISA-based assay and can be pe o med in any ou ine lab a a lowe cos . The ExaVi Load assay equi es a cheape and main enance ee s a -up ki as compa ed o eal- ime assays. We obse ed ha pe o ming he HIV RT assay ou inely in ou cen e would sa e us $20.2 pe sample as compa ed o an HIV-1 RNA assay. A mo e basic labo a o y, which equi es ins alling basic ELISA equipmen , would sa e $19.6 pe sample ( able 3). The use o CD4 cell coun as a p ognos ic ma ke has been deba ed; i is a gued ha his coun may no eflec he ac ual i al load s a us o he pa ien . 38 The cos asso- cia ed wi h i al load moni o ing using HIV-1 RNA assays, despi e being lowe han PCR assays, is a majo limi ing ac o o i s implemen a ion. Cu en ly, in India, i al load es ing has been phased in o suppo pa ien s ailing fi s - line ART. In he yea 2012, abou 4157 i al load es s we e pe o med unde he Na ional AIDS Con ol O ganisa ion (NACO). 39 Conside ing ha $19.6 could ha e been sa ed pe sample by pe o ming an HIV RT assay, he cos sa ing o hese 4157 i al load es s could ha e amoun ed o $ 81 477 i an HIV RT assay had been used. In ou cos -compa ison analysis, we ha e used a p o- ide pe spec i e (lab se ice). We acknowledge ha his is a na owe pe spec i e han a socie al one, which would include pa ien cos s, oppo uni y cos s among o he cos s. Howe e , he pu pose o ou analysis was o p o ide in o ma ion o labo a o ies in esou ce cons ained se ings, o en aced wi h decisions in he ace o igh budge s, and hus a socie al pe spec i e was no conside ed necessa y. A labo a o y manage aced wi h a limi ed budge would concen a e en i ely on cos s ha ha e an immedia e impac on he /his own budge ; his is he pe spec i e adop ed in his s udy. NACO is now conside ing aking up he moni o ing app oach ecommended by he WHO o diagnose and confi m ART ailu e. Conside ing ha he e a e cu - en ly 604 987 HIV-1-in ec ed indi iduals ecei ing fi s - line ART a 380 cen es sp ead ac oss he coun y, he cos educ ion o u ilising HIV RT compa ed o HIV RNA plasma load can be subs an ial. 40 Scaling up ART equi es he c i ical suppo o HIV-1 i al load moni o ing. E idence om he compa a i e pe o mance o he HIV RT assay wi h HIV-1 RNA assays om ou s and o he s udies om India indica es ha he ExaVi Load assay could se e as an a o dable al e - na i e o moni o pa ien s on ART. Au ho a ilia ions 1 Di ision o Clinical Vi ology, Depa men o Mic obiology, S . John’s Medical College, Bangalo e, India 2 Di ision o Clinical Mic obiology, Depa men o Labo a o y Medicine, Ka olinska Ins i u e , S ockholm, Sweden 3 Depa men o In e na ional Heal h, Uni e si y o Tampe e School o Medicine, Tampe e, Finland 4 Depa men o Public Heal h Sciences, Ka olinska Ins i u e, S ockholm, Sweden 5 Ca idi AB, Uppsala, Sweden 6 Depa men o Pedia ics, S John’s Medical College Hospi al, Bangalo e, India Acknowledgemen s The au ho s would like o hank he HIVIND s udy eam o hei ou s anding e o s in ec ui ing and ollowing he s udy pa icipan s and collec ion o samples. They would also like o hank P a a Nalini Sahoo and Shwe ha D Rao, S . John’s Resea ch Ins i u e, o pe o ming labo a o y expe imen s pe aining o HIV-1 d ug esis ance and HIV-1 RNA load. They a e e y much hank ul o S a an Sjödahl, Ca idi AB o his echnical help and suppo wi h he ExaVi Load V3 assay. Ujjwal Neogi acknowledges he suppo ecei ed om Ka olinska Ins i u e Resea ch Founda ion G an s (2014 obi41250). The au ho s also acknowledge he s a a he ART Cen e, S . John’s Medical College Hospi al and KR Hospi al, Myso e o hei excellen eamwo k and pa ien ca e and NACO and Ka na aka S a e AIDS P e en ion Socie y (KSAPS) o hei suppo . Mos impo an ly, hey a e e e hank ul o he pa ien s o hei pa icipa ion in he s udy. Con ibu o s UN, ADC, CK, PA and AS concei ed he s udy. SG, UN, HS, ADC, CK and AS designed he expe imen s. SG and RP pe o med he expe imen s. PA was esponsible o ex e nal lab moni o ing and in e im quali y assessmen . SG, UN, ADC, CK and AS analysed he da a. CK and AS con ibu ed eagen s/ma e ials/analysis ools. SG, RP, UN, HS, PA, ADC, CK and AS con ibu ed o he w i ing o he manusc ip . Funding The s udy was inancially suppo ed by a g an unde he Eu opean Union F amewo k P og amme 7 (p ojec no: 222946). Compe ing in e es s CK is an employee o Ca idi AB, Uppsala, Sweden. All he o he au ho s ha e decla ed ha no compe ing in e es s exis . Pa ien consen Ob ained. E hics app o al E hical app o als o he conduc o he ial we e ob ained om he Ins i u ional E hical Re iew Boa d o S John’s Medical College Hospi al, Bangalo e (IERB 1/369/08-92/2008) and K ishna Rajend a Hospi al, Myso e (NO/PS/173/2010). All pa ien s pa icipa ing in he HIVIND s udy ha e gi en hei w i en consen . Gup a S, e al.BMJ Open 2016;6:e008795. doi:10.1136/bmjopen-2015-008795 7 Open Access g oup.bmj.com on Ap il 20, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om P o enance and pee e iew No commissioned; ex e nally pee e iewed. Da a sha ing s a emen No addi ional da a a e a ailable. Open Access This is an Open Access a icle dis ibu ed in acco dance wi h he C ea i e Commons A ibu ion Non Comme cial (CC BY-NC 4.0) license, which pe mi s o he s o dis ibu e, emix, adap , build upon his wo k non- comme cially, and license hei de i a i e wo ks on di e en e ms, p o ided he o iginal wo k is p ope ly ci ed and he use is non-comme cial. See: h p:// c ea i ecommons.o g/licenses/by-nc/4.0/ REFERENCES 1. Neogi U, Heylen E, She A, e al. Long- e m e icacy o i s line an i e o i al he apy in Indian HIV-1 in ec ed pa ien s: a longi udinal coho s udy. PLoS ONE 2013;8:e55421. 2. An i e o i al The apy Coho Collabo a ion. Li e expec ancy o indi iduals on combina ion an i e o i al he apy in high-income coun ies: a collabo a i e analysis o 14 coho s udies. Lance 2008;372:293–9. 3. B yan L, Smi h N, Keise P. A model o educed HIV-1 i al load moni o ing in esou ce-limi ed se ings. J In Assoc P o id AIDS Ca e 2013;12:67–71. 4. C owe S, Tu nbull S, Oel ichs R, e al. Moni o ing o human immunode iciency i us in ec ion in esou ce-cons ained coun ies. Clin In ec Dis 2003;37(Suppl 1):S25–35. 5. Rawizza HE, Chaplin B, Meloni ST, e al. Immunologic c i e ia a e poo p edic o s o i ologic ou come: implica ions o HIV ea men moni o ing in esou ce-limi ed se ings. Clin In ec Dis 2011;53:1283–90. 6. Wilson D, Keiluhu AK, Kog um S, e al. HIV-1 i al load moni o ing: an oppo uni y o ein o ce ea men adhe ence in a esou ce-limi ed se ing in Thailand. T ans R Soc T op Med Hyg 2009;103:601–6. 7. Iqbal HS, Balak ishnan P, Cecelia AJ, e al. Use o an HIV-1 e e se- ansc ip ase enzyme-ac i i y assay o measu e HIV-1 i al load as a po en ial al e na i e o nucleic acid-based assay o moni o ing an i e o i al he apy in esou ce-limi ed se ings. JMed Mic obiol 2007;56(P 12):1611–14. 8. S ewa P, Cacha ei o A, Nap a nik S, e al. Pe o mance cha ac e is ics o he Ca idi ExaVi i al load assay and he ul a-sensi i e P24 assay ela i e o he Roche Moni o HIV-1 RNA assay. J Clin Vi ol 2010;49:198–204. 9. S e ens WS, Sco LE, C owe SM. Quan i ying HIV o moni o ing an i e o i al he apy in esou ce-poo se ings. J In ec Dis 2010;201 (Suppl 1):S16–26. 10. Si apalasingam S, Wangechi B, Ma shed F, e al. Moni o ing i ologic esponses o an i e o i al he apy in HIV-in ec ed adul s in Kenya: e alua ion o a low-cos i al load assay. PLoS ONE 2009;4: e6828. 11. Seyoum E, Wolday D, Gi ma M, e al. Re e se ansc ip ase ac i i y o quan i a ion o HIV-1 sub ype C in plasma: ela ion o RNA copy numbe and CD4 T-cell coun . J Med Vi ol 2006;78:161–8. 12. Malms en A, Shao XW, Sjodahl S, e al. Imp o ed HIV-1 i al load de e mina ion based on e e se ansc ip ase ac i i y eco e ed om human plasma. J Med Vi ol 2005;76:291–6. 13. Lomba JP, V ay M, Ka ando A, e al. Plasma i ion e e se ansc ip ase ac i i y and hea dissocia ion-boos ed p24 assay o HIV load in Bu kina Faso, Wes A ica. AIDS 2005;19:1273–7. 14. Labbe W, Ga cia-Diaz A, Fox Z, e al. Compa a i e e alua ion o he ExaVi Load e sion 3 e e se ansc ip ase assay o measu emen o human immunode iciency i us ype 1 plasma load. J Clin Mic obiol 2009;47:3266–70. 15. Jennings C, Fiscus SA, C owe SM, e al. Compa ison o wo human immunode iciency i us (HIV) RNA su oga e assays o he s anda d HIV RNA assay. J Clin Mic obiol 2005;43:5950–6. 16. Huang D, Zhuang Y, Zhai S, e al. HIV e e se ansc ip ase ac i i y assay: a easible su oga e o HIV i al load measu emen in China. Diagn Mic obiol In ec Dis 2010;68:208–13. 17. G eeng ass VL, Tu nbull SP, Hocking J, e al. E alua ion o a low cos e e se ansc ip ase assay o plasma HIV-1 i al load moni o ing. Cu HIV Res 2005;3:183–90. 18. G eeng ass VL, Pla e MM, S eele PM, e al. E alua ion o he Ca idi ExaVi Load assay ( e sion 3) o plasma human immunode iciency i us ype 1 load moni o ing. J Clin Mic obiol 2009;47:3011–13. 19. G eeng ass V, Lohman B, Mo is L, e al. Assessmen o he low-cos Ca idi ExaVi Load assay o moni o ing HIV i al load in pedia ic and adul pa ien s. J Acqui Immune De ic Synd 2009;52:387–90. 20. Shao Y, Williamson C. The HIV-1 epidemic: low- o middle-income coun ies. Cold Sp ing Ha b Pe spec Med 2012;2:a007187. 21. Mine M, Bedi K, Ma u a T, e al. Quan i a ion o human immunode iciency i us ype 1 i al load in plasma using e e se ansc ip ase ac i i y assay a a dis ic hospi al labo a o y in Bo swana: a decen aliza ion pilo s udy. J Vi ol Me hods 2009;159:93–7. 22. S e ens G, Rekh iash ili N, Sco LE, e al. E alua ion o wo comme cially a ailable, inexpensi e al e na i e assays used o assessing i al load in a coho o human immunode iciency i us ype 1 sub ype C-in ec ed pa ien s om Sou h A ica. J Clin Mic obiol 2005;43:857–61. 23. She A, Cos a AD, Kuma asamy N, e al. E ec o mobile elephone eminde s on ea men ou come in HIV: e idence om a andomised con olled ial in India. BMJ 2014;347:g5978. 24. NACO, Depa men o AIDS Con ol, Minis y o Heal h and Family Wel a e, Annual Repo 2012–13. h p://www.nacoonline.o g/upload/ Publica ion/AnnualRepo /Annual epo 2012-13_English.pd (accessed on 8 Jul 2013). 25. Neogi U, P a hana BS, Gup a S, e al. Na u ally occu ing polymo phisms and p ima y d ug esis ance p o ile among an i e o i al-nai e indi iduals in Bangalo e, India. AIDS Res Hum Re o i uses 2010;26:1097–101. 26. Gup a S, Neogi U, S ini asa H, e al. HIV-1 co- ecep o opism in India: inc easing p opo ion o X4- opism in sub ype C s ains o e wo decades. J Acqui Immune De ic Synd 2014;65:397–404. 27. Neogi U, Bon ell I, She A, e al. Molecula epidemiology o HIV-1 sub ypes in India: o igin and e olu iona y his o y o he p edominan sub ype C. PLoS ONE 2012;7:e39819. 28. Neogi U, Gup a S, Palchaudhu i R, e al. Limi ed e olu ion bu inc easing ends o p ima y non-nucleoside e e se ansc ip ase inhibi o esis ance mu a ions in he apy-nai e HIV-1-in ec ed indi iduals in India. An i i The (Lond) 2014;19:813–18. 29. Benne DE, Camacho RJ, O elea D, e al. D ug esis ance mu a ions o su eillance o ansmi ed HIV-1 d ug- esis ance: 2009 upda e. PLoS ONE 2009;4:e4724. 30. S eegen K, Luch e s S, De Caboo e N, e al. E alua ion o wo comme cially a ailable al e na i es o HIV-1 i al load es ing in esou ce-limi ed se ings. J Vi ol Me hods 2007;146:178–87. 31. Wo ld Heal h O ganiza ion. Consolida ed guidelines on he use o an i e o i al d ugs o ea ing and p e en ing HIV in ec ion. Recommenda ions o a public heal h app oach. Gene a, Swi ze land: Wo ld Heal h O ganiza ion, 2013. 32. Kokkayil P, Ku apa i S, Negi N, e al. Compa a i e e alua ion o a e e se ansc ip ase based assay o HIV-1 i al load quan i a ion in esou ce limi ed se ings. J Vi ol Me hods 2014;203:1–4. 33. Malms en A, Shao XW, Ape ia K, e al. HIV-1 i al load de e mina ion based on e e se ansc ip ase ac i i y eco e ed om human plasma. J Med Vi ol 2003;71:347–59. 34. an Rooijen LB, G eeng ass V, Mo is LM, e al. E ec o e e se ansc ip ase inhibi o s and mu a ions on he low-cos Ca idi e e se ansc ip ase i al load assay. J Acqui Immune De ic Synd 2009;52:527–9. 35. Nap a nik S, Cacha ei o A, S ewa P, e al. HIV-1 i al load and pheno ypic an i e o i al d ug esis ance assays based on e e se ansc ip ase ac i i y in compa ison o ampli ica ion based HIV-1 RNA and geno ypic assays. J Clin Vi ol 2010;47:18–22. 36. Shao XW, Malms en A, Lenne s and J, e al. Use o HIV-1 e e se ansc ip ase eco e ed om human plasma o pheno ypic d ug suscep ibili y es ing. AIDS 2003;17:1463–71. 37. Al a ez-U ia G, Azcona JM, Reddy S, e al. Poin o ca e es ing o HIV in child en younge han 18 mon hs wi h h ee di e en HIV i ological assays. Expe ience om a dis ic hospi al in a esou ce-limi ed se ing. Mic obiology Resea ch 2012;3:e14. 38. Sax PE. Edi o ial commen a y: can we b eak he habi o ou ine CD4 moni o ing in HIV ca e? Clin In ec Dis 2013;56:1344–6. 39. A ahe Q, Po smou h S, Napoli ano LA, e al. The epidemiology and clinical co ela es o HIV-1 co- ecep o opism in non-sub ype B in ec ions om India, Uganda and Sou h A ica. J In AIDS Soc 2012;15:2. 40. Vi o ia M, Vella S, Fo d N. Scaling up an i e o i al he apy in esou ce-limi ed se ings: adap ing guidance o mee he challenges. Cu Opin HIV AIDS 2013;8:12–18. 8Gup a S, e al.BMJ Open 2016;6:e008795. doi:10.1136/bmjopen-2015-008795 Open Access g oup.bmj.com on Ap il 20, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om moni o ing in esou ce-limi ed se ings? be a low-cos al e na i e o i al load Can HIV e e se ansc ip ase ac i i y assay Asho n, Ayesha De Cos a, Clas Källande and Ani a She Soham Gup a, Riya Palchaudhu i, Ujjwal Neogi, Hi esa e S ini asa, Pe doi: 10.1136/bmjopen-2015-008795 2016 6: BMJ Open h p://bmjopen.bmj.com/con en /6/1/e008795 Upda ed in o ma ion and se ices can be ound a : These include: Re e ences #BIBLh p://bmjopen.bmj.com/con en /6/1/e008795 This a icle ci es 38 a icles, 12 o which you can access o ee a : Open Access h p://c ea i ecommons.o g/licenses/by-nc/4.0/non-comme cial. See: p o ided he o iginal wo k is p ope ly ci ed and he use is non-comme cially, and license hei de i a i e wo ks on di e en e ms, pe mi s o he s o dis ibu e, emix, adap , build upon his wo k Commons A ibu ion Non Comme cial (CC BY-NC 4.0) license, which This is an Open Access a icle dis ibu ed in acco dance wi h he C ea i e se ice Email ale ing box a he op igh co ne o he online a icle. Recei e ee email ale s when new a icles ci e his a icle. Sign up in he Collec ions Topic A icles on simila opics can be ound in he ollowing collec ions (405)In ec ious diseases (137)HIV AIDS No es h p://g oup.bmj.com/g oup/ igh s-licensing/pe missions To eques pe missions go o: h p://jou nals.bmj.com/cgi/ ep in o m To o de ep in s go o: h p://g oup.bmj.com/subsc ibe/ To subsc ibe o BMJ go o: g oup.bmj.com on Ap il 20, 2016 - Published by h p://bmjopen.bmj.com/Downloaded om