Resistance in infants and children receiving ART in South Africa

Polly Clayden, HIV i-Base

Virological outcomes and resistance patterns in children receiving ART in Africa are not well characterised. Two presentations at the Eleventh International Congress on Drug Therapy in HIV Infection showed resistance data from the Children with HIV Early Antiretroviral Therapy (CHER) trial and a cohort in Cape Town respectively.

Avy Violari presented data from CHER. [1] This trial, conducted in South Africa, compared deferred but continuous ART to early limited ART in young infants who were followed for up to six years. Interim data from CHER led to early ART irrespective of CD4 count or disease progression to be recommended in all paediatric guidelines from 2008.

In the trial, all infants received lopinavir/ritonavir (LPV/r) plus AZT and 3TC and were randomised to one of three arms: to start immediately and stop at 40- or 96-weeks and restart when CD4 percent fell below 20% or clinically indicated, or defer ART until clinical progression or CD4 percent drop. Final results were presented at CROI last year and showed early limited ART was safe in children with regular clinical and CD4 monitoring. [2] Only 7 children in switched to second line ART during follow up.

The resistance analysis was conducted to compare rates of virological suppression at last visit on ART and look at resistance at virological failure

Viral load was measured in all children with a stored sample at their last visit, having been on initial or restarted ART, and followed for at least 24 weeks.

At enrollment, infants were a median age of just over 7 weeks, with a CD4 percent of 35% and a viral load of 5.7 log10. Through PMTCT strategies, 63% had been exposed to nevirapine (NVP), 11% were unexposed, 4% had AZT exposure, 20% had AZT plus NVP, and 3% were exposed to maternal ART. The infants receiving deferred ART started at a median age of 26.1 weeks and those in the two immediate arms at 7.5 weeks. The median duration of ART was 240 weeks and similar across all three arms.

Of a total of 377 children, 46 were excluded: 17, 7 and 4 died; and 3, 4 and 5 were lost to follow up in the deferred ART, 40-week ART and 96-week ART arms respectively. In the deferred ART and 40-week ART arms 4 and 2 children never started treatment. The remaining 331 children were included in the resistance analysis.

At the last study visit, viral load was <400 copies/mL in 88/101 (87%), 95/113 (84%) and 97/117 (83%) in the deferred ART, 40- and 96-week arms. Respectively 5/101 (5%), 14 (12%) and 13 (11%) had viral load >1000 copies/mL.

Resistance testing was performed on all but one infant in the 40-week ART arm and, of these, 16/31 (52%) had mutations. Testing revealed no infant with thymidine analogue mutations (TAMs); 11 (35%) with mutations conferring resistance to NRTIs: 10 (33%) with M184V and 1 with L74V; 2 (6%) with major PI mutations (V82A), and 6 (19%) infants with major NNRTI mutations (K101, K103, Y181).

Baseline resistance testing results were available for 21/32 of the infants with last time point mutations. Of these, 8/21 had major NNRTI mutations prior to starting first line ART; 3/8 were also detected at last visit, despite no exposure to NNRTI except through PMTCT. There were no PI or NRTI mutations at baseline.

There were no statistical differences in time to virological failure across the three arms.

In a related presentation, Catherine Orrel showed findings from a retrospective analysis of receiving ART at the Hannan Crusaid Treatment Centre, a public sector ART clinic in Cape Town, between 2003 and 2010.

Children in this cohort were treated with either NNRTI- or LPV/r-based regimens first line, except for infants <6 months old who received full-dose ritonavir-based ART from 2004-2007. Those switching to second line received the alternative regimen.

Resistance testing was performed on stored samples from children at first or second line virological failure, defined as viral load >1000 copies/mL.

Out of 472 children starting first line ART in this cohort, 279 (60%) remained in care, 45 (9%) were lost to follow up, 73 (15%) transferred, and 4 (1%) died on first-line treatment. Seventy-one (15%) children had virological failure and 37 of these had samples available for genotype testing. The median age of the children with genotype results was 5.5 years (IQR 1.6 – 7.7). Eight (22%) children had wild-type virus, 7 (19%) had TAMs, 24 (65%) had NNRTI resistance, and two (5.4%) had multiple PI resistance.

Of 78 children who switched, 48 (63%) remained in care, 6 (8%) were lost, 6 (8%) transferred, and 1 child (1%) died during second-line treatment. Fifteen (20%) had virologic failure and 13 had samples available for testing. The median age of the children with genotype results was 3.6 years (IQR 3.1 – 4.2). Three (23%) had wild-type virus, 8 (62%) had TAMs, nine (69%) had NNRTI resistance, and 5 (38%) had multiple PI resistance (all had received full-dose ritonavir).

Dr Orell noted that a similar proportion of children had wild type virus at first and second-line failure perhaps indicating challenges with adherence. She also noted more TAMs at second-line failure, 19% vs. 62%.

Multiple PI resistance mutations were seen in 50% of those receiving full-dose ritonavir at some time during the analysis but little PI resistance was seen with those receiving LPV/r.


The CHER results confirm the rarity of PI resistance using LPV/r in infants and young children and the ability of a PI to protect against development of TAMs. This was also seen in the arm of the PENPACT1 trial where switching ART was deferred until viral load was 30,000 copies/mL.

CHER also showed that virological suppression was just as good after a treatment interruption and that there was no increase in resistance. Concerns still remain about interrupting treatment in children, based on the negative impact this has in adults.

In the Cape Town cohort, children seemed to do better virologically and develop less resistance wise when started on an NNRTI-based combination and then switched to a PI. However, this is likely to be because those switching from PI were the selected non adherers, more than those moving from NNRTI to PI.

Age may well be a confounding issue here as well as the youngest children would have started on a PI. These data also show little PI resistance in children receiving LPV/r.


  1. Violari A et al. Does early initiation of ART in infants affect virological and resistance outcomes? Data from the CHER trial after 6 years of follow-up. 11th International Congress on Drug Therapy in HIV Infection, Glasgow, 11-15 November 2012. Oral abstract O224.
  2. HTB. Stopping treatment after early ART in infants. April 2012.
  3. Orrell O et al. Evolution of resistance in paediatric patients with failure on antiretroviral therapy. 11th International Congress on Drug Therapy in HIV Infection, Glasgow, 11-15 November 2012.

Links to other websites are current at date of posting but not maintained.