Treating children previously exposed to single dose nevirapine: update on IMPAACT P1060 and NEVEREST

15 August 2011. Related: Conference reports, PMTCT and maternal health, Paediatric care, CROI 18 (Retrovirus) Boston 2011.

Polly Clayden, HIV i-Base

Two oral presentations at CROI 2011 showed further findings from studies looking at treatment in children previously exposed or unexposed to maternal/infant single dose nevirapine (NVP) in prevention of mother to child transmission (PMTCT) programmes.

IMPAACT 1060

IMPAACT P1060 was a randomised trial to determine whether NVP- or lopinavir/ritonavir (LPV/r)-based treatment performed better in young children exposed and unexposed to single dose NVP. All children received AZT plus 3TC. The trial comprised of Cohort 1 (exposed children) and Cohort 2 (unexposed children). Data from Cohort 1 have previously been reported and this part of the study was stopped early after a scheduled Data Safety Monitoring Board (DSMB) review, as there was an unsurprising trend towards more failure in the children receiving NVP- compared to LPV/r-based treatment.

Peter Palumbo presented results from Cohort 2. This cohort enrolled children aged 2 to 36 months, who met WHO criteria for treatment and were unexposed to single dose NVP. Children were stratified by age < or ≥ 12 months. Children with TB were excluded from the trial.

The study had a composite primary endpoint of treatment failure, which comprised viral failure (<1 log10 decline from baseline to after 12 to 24 weeks or >400 copies/mL at week 24), or permanent discontinuation of NVP or LPV/r, including death by 24 weeks. Rates were calculated from Kaplan-Meier curves for each treatment group and age group.

Secondary endpoints included time to virological failure by 24 weeks, time to treatment failure throughout follow up and time to virological failure or death throughout follow up.

P1060 Cohort 2 was fully enrolled with 288 children by March 2010 and had 48-week planned follow-up to March 2011.  In October 2010, the DSMB recommended that the study was unblinded. All children had completed 24 weeks of follow up.

Dr Palumbo reported that the children’s median age at enrollment was 1.7 years (73% >12 months) and their median baseline viral load and CD4 percentage were 535,632 copies/mL and 15% respectively. The majority (79%) of children were subtype C.  The median follow-up was 72 weeks.

At week 24, 87 children had reached an endpoint; 60 in the NVP and 27 in the LPV/r arms. The overall difference in failure rate was 21.5% (95% CI, 11.2-31.8) in favour of LPV/r, p<0.001. This was similar in both age groups: 22.0% (<12 months) and 21.3% (>12 months).

There was also a significant difference in time to off study drug, over the full length of the trial, p<0.001. There were 10 vs 3 deaths in the NVP vs LPV/r arms during the entire follow-up

period (none judged related to study drugs), but this did not reach statistical significance, p=0.63.

There was a notable amendment during the course of the trial. In 2007 the recommended NVP dose in WHO guidelines increased from the FDA recommended dose of 7mg/kg to 160-200mg/m2 (max 200mg). Only 32 children were enrolled under the lower dose compared to 115 at the higher one but the investigators saw no effect associated with this change.

Dr Palumbo noted that the main reasons for off study were more virological failure, toxicity and death in the NVP arm.

As both the NEVEREST and P1060 Cohort 1 data had suggested poorer weight and CD4 improvement in children receiving LPV/r compared to NVP, the investigators also looked at this in Cohort 2. They did not find a statistically significant difference in CD4 improvement between the two arms but there was a difference in weight z-score favouring NVP at 24 and 48 weeks, respectively p=0.007 and p=0.009.

When the investigators looked at NVP resistance in samples from subsets of children at baseline and time of virological failure, they found 2.4% (5/206) with resistance at baseline compared to 56% (10/18) at time of virological failure.

These results were different to those in the sister study, OCTANE P1060, in which maternal data demonstrated non-inferiority of NVP- to LPV/r-based treatment, by the study definition, for NVP- unexposed women.

This highlighted the “unique and challenging situation of early paediatric HIV infection”, Dr Palumbo said, including very high baseline viral load and the unforgiving nature of NVP resistance. LPV/r is already recommended for NVP-exposed children and discussions are ongoing as to whether this recommendation should expand to all young children, possibly up to three years of age.

These data once again point to the importance of developing new first and second line options for use in this age group.

NEVEREST

Louise Kuhn presented data from NEVEREST, a study designed to evaluate a treatment switch strategy from LPV/r to NVP in NVP-exposed children.

In this study, 323 children aged 6 weeks to 2 years and eligible for treatment were initiated on LPV/r plus 3TC plus d4T. After achieving a viral load <400 copies/mL and maintaining it for > 3 months, children were randomised (n=195) to either remain on LPV/r (n=99) or switch to NVP (n=96). Time to any viral load >50 copies/mL or confirmed >1000 copies/mL was compared using Kaplan-Meier methods and log-rank tests.

Fifty-two week data post switch from this study has been reported previously. These data revealed a higher proportion of children suppressed to <50 copies/mL (the primary endpoint) in the NVP arm but also a higher proportion in that group with confirmed >1000 copies/mL.

Dr Kuhn showed longer term results from this study with follow up of 18-53 months.

There were three deaths in each group. At 36 months post randomisation, as with the earlier analysis, more children in the NVP group (40.5%) maintained viral load <50 copies/mL than those in the LPV/r group, p=0.01. Again, more in the NVP (23.9%) than in the LPV/r (11.1%) had confirmed >1000 copies/mL, p=0.01.

This difference persisted at 48 months, for <50 copies/mL and >1000 copies/mL, respectively p=0.02 and p=0.08.

At 6 months 59.1% of the failures in the NVP group had occurred vs 10% in the LPV/r group. By 12 months these proportions were 100% in the NVP group and 50% in the LPV/r group. Dr Kuhn noted that among children in the LPV/r group, 6% of failures occurred between 12 and 48 months.

Treatment failure >1000 copies/mL was associated with the presence of pre-treatment NVP mutations, p=0.02. There was no difference in response between children in the NVP and LPV/r groups in children who had no pre-treatment NVP resistance. Half the children with detectable NVP mutations failed when re-challenged with NVP.

Dr Kuhn concluded that viral load testing can identify all switch failures and that switching can be accomplished safely if viral load testing is available. Also that pre-treatment screening for resistance can be used to identify the children who could benefit from this strategy.

References

1. Palumbo P et al. NVP- vs LPV/r-based ART among HIV⁺ infants in resource-limited settings: the IMPAACT P1060 Trial. 18th CROI, 27 February–2 March 2011, Boston. Oral abstract 129LB.
2. Kuhn L et al. Long-term outcomes of switching children to NVP-based therapy after initial suppression with a PI based regimen. 18th CROI, 27 February–2 March 2011, Boston. Oral abstract 128.