You are here
TIME TO VIRAL REBOUND AFTER STOPPING ART IN CHILDREN TREATED FROM INFANCY IN CHER
Avy Violari1, Man Chan2, Kennedy N. Otwombe1, Ravindre Panchia1, Patrick Jean-Philippe3, Diana Gibb2, Mark Cotton4, Abdel Babiker2
1Perinatal HIV Research Unit, Soweto, South Africa,2University College London, London, UK,3NIAID, Bethesda, MD, USA,4Stellenbosch University, Cape Town, South Africa
We investigated factors associated with time to viral rebound in children in CHER who started ART at age <12 weeks and received 40 (ART-40W) or 96 weeks (ART-96W) of primary therapy.
HIV RNA viral load (VL) from stored samples was assessed 8 weeks after interruption and 12 weekly thereafter. Included were children with VL<400 c/ml at interruption and ≥1 VL measurement within 12 months. Multivariable stepwise Cox regression models (backwards elimination, exit probability p=0.05) were used to identify factors associated with time to viral rebound (confirmed VL≥400 c/ml). Follow-up was censored at ART reinitiation (if VL had not rebounded) or last VL measurement.
Of 183 children virally suppressed (VL<400) at interruption, 54% were from ART-40W and 61% were female. At enrolment, 81% received PMTCT, 81% had CDC stage N; median [IQR] birth weight was 3 [2.7,3.3]Kg. At ART start, median [IQR] age was 1.8 [1.5,2.1] months, CD4% 34 [29,40]%, CD4 count 1982 [1445,2745], CD8% 28 [22,34]% and VL 750000 [376000,750000] copies/ml. Median VL at rebound was 354615 [91040,750000] copies/ml, not significantly different between arms [ART-40W=418760; ART-96W=325000 copies/ml; P=0.19]. 86% of children suppressed within 40 weeks of ART start [88% ART-40W; 83% ART-96W; P=0.38]. Overall estimated cumulative probability of rebound (95% CI) at 2, 4, 6 and 8 months were 70% (63,76)%, 80% (74, 85)%, 94% (90,97)% and 99% (96,100)%, respectively. Median time to rebound was 1.8 (range: 0.9-13.1) months. One child (ART-40W) maintained viral suppression until last VL available. Five children were censored due to ART restart. In univariable analysis, among baseline demographic and clinical factors, CD4% was the strongest predictor of longer time to rebound based on the log likelihood ratio. In multivariable analysis, longer time to rebound was associated with higher birth weight, baseline CD4% and viral suppression within 40 weeks of ART start (Table). There was no evidence of significant effect of gender, baseline VL and CD8%, CDC stage, PMTCT, age at ART initiation (6-12weeks) and length of therapy (arm) or site. Sensitivity analyses produced similar results.
Most children rebounded by 13 months while one remained suppressed until the end of follow up. Age at ART initiation ranging from 6 to 12 weeks and length of therapy were not associated with longer time to rebound. Our findings may inform the design of clinical trials involving analytic treatment interruption in paediatric HIV.