Height and timing of growth spurt during puberty in young people living with vertically acquired HIV in Europe and Thailand

Heigh and iming o g ow h spu du ing pube y in young people li ing wi h e ically acqui ed HIV in Eu ope and Thailand The Eu opean P egnancy and Paedia ic HIV Coho Collabo a ion (EPPICC) s udy g oup M Objec i e: The aim o his s udy was o desc ibe g ow h du ing pube y in young people wi h e ically acqui ed HIV. Design: Pooled da a om 12 paedia ic HIV coho s in Eu ope and Thailand. Me hods: One housand and nine y- ou child en ini ia ing a nonnucleoside e e se ansc ip ase inhibi o o boos ed p o ease inhibi o based egimen aged 1–10 yea s we e included. Supe Imposi ion by T ansla ion And Ro a ion (SITAR) models desc ibed g ow h om age 8 yea s using h ee pa ame e s (a e age heigh , iming and shape o he g ow h spu ), dependen on age and heigh - o -age z-sco e (HAZ) (WHO e e ences) a an i e o i al he apy (ART) ini ia ion. Mul i a ia e eg ession explo ed cha ac e is ics associa ed wi h hese h ee pa ame e s. Resul s: A ART ini ia ion, median age and HAZ was 6.4 [in e qua ile ange (IQR): 2.8, 9.0] yea s and 1.2 (IQR: 2.3 o 0.2), espec i ely. Median ollow-up was 9.1 (IQR: 6.9, 11.4) yea s. In gi ls, olde age and lowe HAZ a ART ini ia ion we e independen ly associa ed wi h a g ow h spu which occu ed 0.41 (95% con idence in e al 0.20– 0.62) yea s la e in child en s a ing ART age 6 o 10 yea s compa ed wi h 1 o 2 yea s and 1.50 (1.21–1.78) yea s la e in hose s a ing wi h HAZ less han 3 compa ed wi h HAZ a leas 1. La e g ow h spu s in gi ls esul ed in con inued heigh g ow h in o la e adolescence. In boys s a ing ART wi h HAZ less han 1, g ow h spu s we e la e in child en s a ing ART in he oldes age g oup, bu o HAZ a leas 1, he e was no associa ion wi h age. Gi ls and boys who ini ia ed ART wi h HAZ a leas 1 main ained a simila heigh o he WHO e e ence mean. Conclusion: S un ing a ART ini ia ion was associa ed wi h la e g ow h spu s in gi ls. Child en wi h HAZ a leas 1 a ART ini ia ion g ew in heigh a he le el expec ed in HIV nega i e child en o a compa able age. Copy igh ß2019 The Au ho (s). Published by Wol e s Kluwe Heal h, Inc. AIDS 2019, 33:1897–1910 Keywo ds: Eu ope, g ow h, heigh , HIV, pe ina al, pube y, Thailand In oduc ion Al hough young people li ing wi h HIVa e a isk o poo heigh g ow h [1], ea men wi h an i e o i al he apy (ART) imp o es g ow h, wi h s onges gains in hose ea ed a a young age [2]. Al hough ini ial ca ch-up g ow h on ART has been well desc ibed [2], he e a e less da a on long e m g ow h, pa icula ly du ing adolescence. Delays in pube al de elopmen ha e been epo ed in young people wi h HIV [3–7], wi h he onse o pube y [5] and sexual ma u a ion [6] occu ing 6 mon hs la e compa ed wi h HIV-exposed unin ec ed young people (HEU). Ea lie pube y in he gene al popula ion is associa ed wi h being alle and ha ing highe BMI h oughou childhood [8], and poo g ow h in child en wi h HIV has been shown o accoun o much o he Co espondence o Siobhan C ich on, PhD, MRC Clinical T ials Uni a UCL, 90 High Holbo n, London WC1 V 6LJ, UK. Tel: +44 0 20 7670 4913; e-mail: [email p o ec ed]  Please see Acknowledgemen s sec ion o lis o W i ing G oup membe s. Recei ed: 6 No embe 2018; e ised: 15 May 2019; accep ed: 17 May 2019. DOI:10.1097/QAD.0000000000002294 ISSN 0269-9370 Copy igh Q2019 The Au ho (s). Published by Wol e s Kluwe Heal h, Inc. This is an open access a icle dis ibu ed unde he C ea i e Commons A ibu ion License 4.0 (CCBY), which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed. 1897 delay in eaching sexual ma u i y [6]. The e is also e idence ha child en s a ing ARTwi h low heigh - o - age z-sco es expe ience delays in he onse o pube y independen ly o age a ART ini ia ion [3]. Poo g ow h du ing childhood can ha e implica ions o u u e heal h. Heigh eloci y is associa ed wi h inc eased HIV eplica ion [9] and p og ession o AIDS and dea h [10] wi h he associa ion wi h dea h being independen o age, i al load and CD4 þ cell coun [11]. The iming o pube y is also in e sely associa ed wi h bone mass and densi y among HIV-nega i e adolescen s [12] and delayed pube y may inc ease u u e isk o os eopo osis among young people wi h HIV, who hemsel es a e a isk o poo bone heal h, ei he caused by HIV in ec ion i sel o p olonged exposu e o ART [13]. Ea ly g ow h ailu e has also been linked o poo e social and economic ou comes in la e li e in he gene al popula ion [14]. In his s udy, s a is ical models ha desc ibe an indi idual’s g ow h in e ms o mean heigh h oughou adolescence, and iming and shape o he adolescen g ow h spu we e applied o longi udinal heigh measu emen s. The o e all aim o his s udy was o explo e he associa ion be ween cha ac e is ics a ART ini ia ion, in pa icula age and heigh - o -age z-sco e, and g ow h du ing adolescence. Ma e ials and me hods Se en een paedia ic HIV coho s om 15 coun ies con ibu ed indi idual le el da a o he Eu opean P egnancy and Paedia ic HIV Coho Collabo a ion (EPPICC) be ween Sep embe 2016 and Ma ch 2017 using a modi ied HICDEP p o ocol (www.hicdep.o g). Pseudo-anonymized da a on all child en a pa icipa ing clinics we e included. All coho s ecei ed app o al om local and/o na ional e hical commi ees. Fi e coho s om h ee coun ies (I aly, Uk aine and h ee om Russia) whe e heigh da a we e no ou inely collec ed (each wi h <20% o child en ha ing a heigh measu e- men a ART ini ia ion) we e excluded. Child en om he emaining 12 coho s we e eligible p o ided hey ini ia ed ART wi h a leas wo nucleoside e e se ansc ip ase inhibi o s (NRTIs) along wi h a nonnucleo- side e e se ansc ip ase inhibi o (NNRTI) o boos ed p o ease inhibi o (bPI); we e 1–10 yea s old a ART ini ia ion; no known o ha e ho izon ally acqui ed HIV; and aged a leas 8 yea s a he end o ollow-up. We excluded child en ini ia ing ART a e age 11 yea s. Fo hose ini ia ing ART a an olde age, i would be di icul o dis inguish be ween changes in g ow h occu ing as a esul o a pube al g ow h spu and as a esul o ini ia ing ART. Child en wi h no heigh eco ded a ART ini ia ion and/o a e 8 yea s o age we e excluded. Heigh measu emen s we e censo ed a he ea lies o 19 h bi hday, ans e o adul ca e, dea h o loss o ollow-up. Heigh and BMI we e con e ed o heigh - o -age z-sco es (HAZ) and BMI- o -age z-sco es (zBMI), using he WHO G ow h S anda d o measu e- men s when child en we e aged unde 5 yea s [15] and he WHO 2007 g ow h e e ence when aged 5–18 yea s [15,16]. Da a checks we e ca ied ou o de ec implausible changes in heigh and/o HAZ. HAZ was ca ego ized acco ding o WHO de ini ions as less han 3 SD (se e e s un ing); 3 o less han 2SD (s un ing); 2 o less han 1 SD; and a leas 1SD. zBMI was ca ego ized as less han 2 SD (unde weigh ); 2 o 1 SD (no mal); mo e han 1 o 2 SD (o e weigh ); and mo e han 2 SD (obese). HAZ and zBMI nea es o ART ini ia ion (closes wi hin 6 mon hs be o e o 1 mon h a e ) we e conside ed baseline measu emen s. O he a iables included we e sex, coun y (Thailand, UK/I eland, Res o Eu ope), age a ART ini ia ion (1– 2, 3–5 o 6–10 yea s), ini ial ART egimen (p o ease inhibi o , NNRTI), being bo n ou side he coun y o he coho (’bo n ab oad’) and immunode iciency a ART ini ia ion classi ied using he WHO immunological classi ica ion: none (CD4% >35%, >30%, >25% o CD4 þ cell coun >500 cells/ml in child en <1, 1–2, 3–4 and 5 yea s, espec i ely), mild (CD4% 30–35%, 25– 30%, 20–25% o CD4 þ cell coun 350–499 cells/ml), ad anced (CD4% 25–29%, 20–24%, 15–19% o CD4 þ cell coun 200–349 cells/ml) o se e e (CD4% <25%, <20%, 15–19% o in child en >5 CD4 þ cell coun <200 cells/ml o CD4% <15%) [17]. S a is ical analysis Cha ac e is ics a ART ini ia ion we e summa ized by HAZ ca ego y. Mean heigh a age 16 yea s was summa ized by age and HAZ a ART ini ia ion and compa ed wi h he WHO e e ence heigh o quan i y di e ences in heigh ollowing he g ow h spu . I was no possible o assess di e ences in inal heigh , as many adolescen s ans e o adul ca e om age 16 yea s, ending ollow-up in EPPICC. Heigh was modelled using Supe Imposi ion by T ansla ion And Ro a ion (SITAR) models [18]. SITAR was de eloped o model g ow h du ing childhood and adolescence and quan i ies di e ences in g ow h ia h ee pa ame e s ep esen ing he iming and shape o he adolescen g ow h spu , as well as a e age heigh . The models can explain up o 99% o he a ia ion be ween indi iduals’ g ow h [18] and can be summa ized as: yi ¼aiþh bi expci  whe e he ou come y i is he heigh o indi idual ia age and h( ) is a na u al cubic spline o heigh o e age. The 1898 AIDS 2019, Vol 33 No 12 pa ame e s ai,biand cia e pa icipan speci ic andom e ec s. ai ep esen s a e age heigh h oughou adoles- cence; nega i e alues indica e sho e heigh o e all. bi ep esen s iming o he pube al g ow h spu ; nega i e alues indica e ea lie pube y. ci ep esen s g ow h eloci y, o he shape o he g ow h spu ; posi i e alues indica e sho e g ow h spu s and a s eepe g ow h eloci y cu e, while nega i e alues indica e he g ow h spu occu s o e a longe du a ion. Co esponding g ow h eloci y cu es can also be es ima ed as he i s de i a i e o he modelled g ow h (heigh ) cu e. Age a peak heigh eloci y (APHV) is co ela ed wi h iming o pube y and o en used as a p oxy o iming o ma u a ion. I commonly occu s in gi ls in Tanne s age 2 o 3 and in Tanne s age 3 o 4 o boys [19,20], hough he e is a ia ion in iming ac oss Tanne s ages [19]. Di e ences in he iming o he g ow h spu es ima ed using SITAR models ha e been shown o be highly co ela ed wi h APHV [18]. All heigh measu emen s (in cm) om age 8 (o s a o ART i a e 8 h bi hday) o 18 yea s we e included. Age and HAZ a ART ini ia ion we e added o he SITAR model as ixed e ec s ha could in luence he mean o a, b and c. Thus, he es ima ed andom e ec s ai,biand ci ep esen he indi idual di e ences in a e age heigh , iming and shape o he g ow h spu no associa ed wi h di e ences in age o heigh a ART ini ia ion. Models we e i ed sepa a ely o boys and gi ls using a spline wi h 6 deg ees o eedom. Log ans o ma ions o bo h age and heigh [18] we e conside ed, bu he un ans o med da a p o ided he bes i . In e ac ions be ween baseline heigh and age we e added whe e app op ia e [model compa ison ca ied ou using Bayes In o ma ion C i e ia (BIC)]. To explo e o he ac o s (sex, coun y, ini ial ART egimen, WHO immunological classi ica ion, zBMI a ART ini ia ion) associa ed wi h g ow h a e allowing o di e ences in baseline age and heigh , he es ima ed ai,bi and ci andom e ec s om he SITAR modelwe e analysed using mul i a iable linea eg ession. In e ac ions be ween each o he ac o s and sex and be ween immunological classi ica ion and HAZ and age a ART ini ia ion we e conside ed. A second model was i ed including zBMI a age 8 yea s ins ead o a ART ini ia ion. Modelling was epea ed in coun ies whe e mo e han 5% o child en we e bo n ab oad and mo e han 5% bo n in he coun y (UK and I eland, Spain and Ne he lands) o explo e di e ences be ween hose bo n ab oad and hose bo n in he coho coun y. Th ee sensi i i y analyses we e ca ied ou : in he i s sepa a e models we e i ed o child en om Thailand and elsewhe e; in he second Thai-speci ic g ow h e e ence da a we e used o Thai child en [21]; and in he hi d child en s a ing ART a e hei eigh h bi hday we e excluded. Analyses we e ca ied ou using S a a s a is ical so wa e elease 15 (S a aCo p LP, College S a ion, Texas, USA) and he SITAR package [22] in R 3.3.3 (R Co e Team, R Founda ion o S a is ical Compu ing, Vienna, Aus ia). Resul s Pa ien cha ac e is ics In o al, 1943 young people wi h HIV ini ia ed ART on an eligible egimen age 1–10 yea s and we e a leas 8 yea s old a he end o ollow-up (Fig. 1). A e excluding hose wi h missing baseline heigh (n¼721) and/o heigh a e age 8 yea s (n¼202), we included 1094 child en in he analysis. Child en excluded due o missing heigh da a we e mo e likely o be om coun ies o he han Thailand o UK/I eland, be bo n ab oad and be younge a ART ini ia ion han hose who we e included (Supplemen a y Table 1, h p://links.lww.com/QAD/ B501). The 1094 included child en we e ollowed-up o a median o 9.1 (6.9, 11.4) yea s a e ART ini ia ion. Du ing his ime, 37 325 heigh measu emen s we e eco ded wi h a median o 32 (19, 46) pe child, o which 25 458 [median 21 (11, 32)] we e om age 8 yea s onwa ds. The median ime be ween heigh measu e- men s was 2.8 (1.4, 3.9) mon hs, wi h some a ia ion by coho anging om e e y 1.8 (1.0, 2.8) mon hs in Thailand o 8.3 (4.7, 11.3) mon hs in G eece. A ART ini ia ion, median HAZ was 1.2 (2.3, 0.2) and age was 6.4 (2.8, 9.0) yea s. Cha ac e is ics o child en a ART ini ia ion, s a i ied by baseline HAZ, a e desc ibed in Table 1. Mo e se e e s un ing was associa ed wi h esidence in Thailand, no being bo n ab oad, ini ia ing on an NNRTI based egimen, ea lie calenda yea o ART ini ia ion, highe i al load, mo e se e e immunode iciency and lowe zBMI a ART ini ia ion. A he end o he s udy, 493 (45%) child en had eached hei 16 h bi hday while s ill in paedia ic ca e (Fig. 1), o whom 463 (94%) had hei heigh eco ded wi hin 6 mon hs o hei bi hday. Child en who su i ed o age 16 yea s bu we e no longe in ollow-up in paedia ic ca e we e mo e likely o eside in Thailand and s a ART a a younge age. A age 16 yea s, he mean (s anda d de ia ion) heigh s o boys and gi ls we e 166 (8.7)cm and 158 (6.9)cm, espec i ely, signi ican ly sho e han he WHO e e ence mean heigh o 173 (7.8)cm o boys and 163 (6.8)cm o gi ls (bo h P<0.001) (Supplemen a y Table 2, h p://links.lww.com/QAD/B501). Associa ions be ween age and heigh - o -age z-sco e a an i e o i al he apy ini ia ion and g ow h om age 8 yea s Resul s om he SITAR models a e a ailable in Supplemen a y Table 3, h p://links.lww.com/QAD/ Pube al g ow h in young people wi h HIV C ich on e al. 1899 B501. Es ima ed mean heigh and co esponding g ow h eloci y cu es s a i ied by HAZ and by age a e summa ized in Fig. 2a and b, espec i ely, o gi ls and Fig. 3a and b o boys. In gi ls, ac oss each o he baseline HAZ g oups (Fig. 2ai– i ), child en s a ing ART in he oldes age g oup had g ow h spu s on a e age 0.41 [95% con idence in e al (95% CI) 0.20–0.62] yea s la e han hose s a ing ART in he younges age g oup. Ac oss he baseline age g oups (Fig. 2bi–iii), gi ls s a ing ARTwi h low HAZ had la e g ow h spu s; he e was a 1.50 (1.21–1.78) yea delay in hose wi h baseline HAZ less han 3 compa ed wi h baseline HAZ a leas 1. The e ec o his delay on o e all heigh can be seen in Fig. 2b i - i; he di e ences in heigh a e smalle om age 16 onwa ds (a e he g ow h spu ) han a age 8 yea s. In boys, he associa ion be ween baseline age and he iming o he g ow h spu di e ed by baseline HAZ (Fig. 3ai–i ); he e was no signi ican di e ence by age in boys who s a ed ARTwi h HAZ a leas 1 (Fig. 3ai). In boys wi h baseline HAZ o 2 o less han 1 (Fig. 3aii), he g ow h spu was 0.96 (0.19–1.72) yea s la e in hose s a ing ART in he oldes compa ed wi h he younges age g oup. Simila ly, o a baseline HAZ o 3 o less han 2 (Fig. 3aiii), he co esponding delay in hose s a ing ART in he oldes age g oup was 0.92 (0.17–1.66) yea s, and o baseline HAZ less han 3, i was 0.42 (0.32 o 1.16) yea s (Fig. 3ai ). The iming o he g ow h spu in boys did no di e signi ican ly by baseline HAZ (Fig. 3bi–iii). Gi ls (Fig. 2b ) and boys (Fig. 3b ), who s a ed ea men wi h a baseline HAZ a leas 1, main ained a simila mean heigh o he WHO e e ence, ega dless o baseline age. O he ac o s associa ed wi h g ow h om age 8 yea s Cha ac e is ics associa ed wi h a ia ions in g ow h ha emained a e adjus ing o di e ences in baseline HAZ and age a e summa ized in Table 2. Young people om Thailand we e smalle h oughou adolescence han hose om o he coun ies, bu did no di e in he iming o he g ow h spu . The shape o he g ow h spu di e ed by coun y and was sho e in child en om he UK and I eland han elsewhe e. Lowe zBMI a ART ini ia ion was 1900 AIDS 2019, Vol 33 No 12 Fig. 1. Flow cha o pa icipan s included in he s udy. signi ican ly associa ed wi h a la e g ow h spu [a one SD dec ease was associa ed wi h a 0.07 (0.02–0.11) yea delay in he g ow h spu ]. In a second model (da a no shown), a one SD dec ease in zBMI a age 8 yea s was associa ed wi h a 0.16 (0.09–0.22) yea delay in he iming o he g ow h spu , while o he pa ame e s did no change subs an ially. The e was no e idence o any in e ac ions. In subg oup analysis (n¼545), he e was a signi ican in e ac ion be ween sex and being bo n ab oad on iming o he g ow h spu (P¼0.038). Gi ls bo n ab oad expe ienced a g ow h spu 0.24 (0.02–0.46) yea s ea lie han hose bo n in he coho coun y, al hough he e was no associa ion in boys. Howe e , a e adjus ing o zBMI a age 8, he associa ion was no longe signi ican [g ow h spu o gi ls bo n ab oad was 0.18 (0.05 o 0.42) yea s ea lie ]. In he h ee sensi i i y analyses whe ein models we e i ed sepa a ely o child en om Thailand and elsewhe e, Thai-speci ic e e ence da a we e used o Thai child en and child en s a ing ART age a leas 8 yea s we e excluded, o e all conclusions we e unchanged (da a no shown). Pube al g ow h in young people wi h HIV C ich on e al. 1901 Table 1. Cha ac e is ics o 1094 young people li ing wi h HIV a an i e o i al he apy ini ia ion s a i ied by heigh - o age z-sco es. Heigh - o -age z-sco e a ART ini ia ion N(%) o median (IQR) All <3SD (Se e ely s un ed) 3 o<2SD (S un ed) 2 o <1SD 1SD P All 1094 157 (14) 187 (17) 272 (25) 478 (44) Male 526 (48) 83 (53) 90 (48) 117 (43) 236 (49) 0.207 Coun y UK and I eland 517 (47) 17 (11) 56 (30) 137 (50) 307 (64) <0.001 Thailand 352 (32) 121 (77) 103 (55) 84 (31) 44 (9) O he a 225 (21) 19 (12) 28 (15) 51 (19) 127 (27) E hnici y Whi e 99 (9) 10 (6) 12 (6) 26 (10) 51 (11) <0.001 Black 484 (44) 16 (10) 49 (26) 125 (46) 294 (62) Asian 365 (33) 121 (77) 106 (57) 89 (33) 49 (10) O he 63 (6) 6 (4) 5 (3) 19 (7) 33 (7) Unknown/P ohibi ed 83 (8) 4 (3) 15 (8) 13 (5) 51 (11) Bo n ab oad 370 (35) 21 (14) 51 (28) 100 (37) 198 (43) <0.001 Age Median yea s 6.4 (2.8–9.0) 6.4 (2.7–9.0) 7.1 (3.8–9.5) 6.4 (3.2–8.7) 6.1 (2.5–8.8) <0.001 1 o 2 yea s 280 (26) 45 (29) 35 (19) 65 (24) 135 (28) 3 o 5 yea s 237 (22) 31 (20) 43 (23) 64 (24) 99 (21) 6 o 10 yea s 577 (53) 81 (52) 109 (58) 143 (53) 244 (51) Yea s a ed ART Median (IQR) 2004 (2003–2007) 2003 (2003–2008) 2004 (200–2006) 2004 (2003–2006) 2005 (2003–2008) <0.001 <2004 433 (40) 79 (50) 78 (42) 116 (43) 160 (33) 2004–2007 445 (41) 69 (44) 93 (50) 106 (39) 177 (37) 2008 216 (20) 9 (6) 16 (9) 50 (18) 141 (30) NNRTI-based egimen 880 (80) 141 (90) 167 (89) 222 (82) 350 (73) <0.001 Vi al load Value p esen 980 (90) 133 (85) 167 (89) 241 (89) 439 (92) Median log VL 5.0 (4.5–5.5) 5.3 (4.9–5.7) 5.0 (4.7–5.5) 5.0 (4.5–5.4) 5.0 (4.2–5.5) <0.001 400 36 (4) 4 (3) 4 (2) 6 (2) 22 (5) >400 o 1000 13 (1) 1 (1) 1 (1) 1 (0) 10 (2) >1000–10 000 91 (9) 7 (5) 11 (7) 21 (9) 52 (12) >10000–100 000 435 (35) 39 (29) 60 (36) 89 (27) 157 (36) >100 000 495 (51) 82 (62) 91 (54) 124 (51) 198 (45) WHO immunological classi ica ion Value p esen 1006 (92) 145 (92) 177 (95) 258 (95) 426 (89) None o no signi ican 164 (16) 7 (5) 16 (9) 40 (16) 101 (24) <0.001 Mild 110 (11) 5 (3) 12 (7) 27 (10) 66 (15) Ad anced 129 (13) 6 (4) 12 (7) 35 (14) 76 (18) Se e e 603 (60) 127 (88) 137 (77) 156 (60) 183 (43) zBMI Value p esen 1089 (100) 157 (100) 186 (100) 271 (100) 475 (99) Median zBMI 0.1 (1.1 o 0.8) 1.1 (2.3 o 0.2) 0.7 (1.6 o 0.3) 0.1 (0.9 o 0.7) 0.3 (0.5 o 1.1) <0.001 <2 SD (Thinness) 133 (12) 49 (31) 34 (18) 22 (8) 28 (6) 2 o<1 SD 164 (15) 32 (20) 42 (23) 42 (16) 48 (10) 1 o<1 SD 576 (53) 59 (38) 96 (62) 158 (58) 263 (55) 1 o<2 SD (O e weigh ) 160 (15) 13 (8) 12 (6) 37 (14) 98 (21) 2 SD (Obese) 56 (5) 4 (3) 2 (1) 12 (4) 38 (8) IQR, in e qua ile ange; VL, i al load; zBMI, BMI- o -age z-sco es. a O he includes Belgium, G eece, La ia, Ne he lands, Poland, Romania, Spain, Sweden and Swi ze land. Discussion In his s udy, we desc ibed g ow h h oughou adoles- cence in a la ge coho o young people wi h e ically acqui ed HIV in Eu ope and Thailand. Al hough all adolescen s in he s udy ini ia ed ART be o e age 11 yea s, g ow h de ici s emained h oughou adolescence. Only child en wi h HAZ a leas 1 when s a ing ART 1902 AIDS 2019, Vol 33 No 12 Fig. 2. (a) Mean heigh and g ow h eloci y o gi ls s a i ied by HAZ a ART ini ia ion and (b) mean heigh and g ow h eloci y o gi ls s a i ied by age a ART ini ia ion. we e able o achie e a simila heigh o he WHO e e ence a age 16 yea s, sugges ing ha o o he s, ca ch up g ow h associa ed wi h being on ART long e m was no su icien o es o e heigh o wha would be expec ed in an HIV-nega i e popula ion. We obse ed an associa ion be ween olde age a ART ini ia ion and la e g ow h spu s in boys (wi h HAZ <1 a ART ini ia ion) and gi ls, in line wi h indings om he An i e o i al esea ch o Wa o o (ARROW) ial whe ein a ainmen o each anne s age and onse o mena che was delayed in hose s a ing ARTa olde ages [3]. We also obse ed an associa ion be ween s un ing and la e g ow h spu s, bu only in gi ls. The po en ial ole o an h opome ic pa ame e s in ea ly childhood on g ow h du ing pube y was highligh ed in a s udy o 2539 young people wi h e ically acqui ed HIV and HIV-exposed unin ec ed (HEU) young people om he USA [6]. Young people li ing wi h HIV eached sexual ma u i y on a e age 6 mon hs la e han he HEU g oup, bu di e ences in HAZ p io o pube y accoun ed o up o 98% o he delay in boys and ( oge he wi h zBMI) 74% in gi ls, sugges ing much o he delay may be a ibu able o ea lie poo g ow h [6]. Low HAZ a ART ini ia ion was also associa ed wi h delayed a ainmen o all Tanne s ages in boys and gi ls, and mena che in gi ls, independen ly o age a ART ini ia ion in he ARROW ial [3]. Howe e , in boys, he delay was educed in hose who had he g ea es ini ial gains in CD4 þ cell coun a e s a ing ART, bu he e was no simila associa ion in gi ls. Unde nu i ion ea ly in li e was also ound o ha e a s onge associa ion wi h adul heigh in women han Pube al g ow h in young people wi h HIV C ich on e al. 1903 Fig. 2. (Con inued). men in he Ne he lands [23]. Al hough his sugges s ha gi ls may be mo e sensi i e o impai men s ea ly in li e, and p io o ART, he mechanism unde lying po en ial sex di e ences emain o be explained. A e accoun ing o HAZ and age a ART ini ia ion, we ound no associa ion be ween WHO immunological s a us o i al load a ART ini ia ion and g ow h. Simila ly, he ARROW ial ound immune supp ession 1904 AIDS 2019, Vol 33 No 12 Fig. 3. (a) Mean heigh and g ow h eloci y o boys s a i ied by HAZ a ART ini ia ion and (b) mean heigh and g ow h eloci y o boys s a i ied by age a ART ini ia ion. p io o ARTwas no associa ed wi h delayed pube y o mena che [3]. O he s udies ha e also epo ed a lack o associa ion be ween clinical s a us a s a o pube y and age a onse [4,7]. Howe e , in young people in he USA, low CD4 þ cell coun and high i al load a i s pube al assessmen we e associa ed wi h la e pube al onse . Among boys, p io CDC C, low nadi CD4% o high peak i al load we e also associa ed wi h la e pube y [5]. Howe e , many o hese young people ini ia ed ART on mono o dual he apy and a e likely o ha e subs an ially di e en ea men his o ies compa ed wi h ou s udy. We ound zBMI a ART ini ia ion and age 8 o be associa ed wi h he iming o he g ow h spu , wi h no e idence o a di e ence be ween boys and gi ls. We also obse ed ha gi ls bo n ab oad expe ienced an ea lie pube al g ow h spu han gi ls bo n in he coho coun y, bu he di e ences in he iming o he g ow h spu educed a e adjus ing o zBMI a age 8. In gi ls, a ela ionship be ween low BMI and delayed pube y has been ound in mul iple s udies [8] and apid weigh gain p io o pube y also linked o ea ly onse [24]. Di e ences be ween young people bo n ab oad and hose bo n in he coun y may he e o e be explained by pe iods o mo e apid weigh gain in child en a i ing om ab oad, he majo i y om A ica, compa ed wi h hose bo n in he coun y. This s udy had se e al limi a ions; as wi h all obse a ional s udies, ou indings on he associa ion be ween age and Pube al g ow h in young people wi h HIV C ich on e al. 1905 Fig. 3. (Con inued).