ICTM logo
Treating infants with HIV early, and then interrupting their treatment, leads to better results than deferring initial treatment

11 November 2013

Treating infants with HIV early, and then giving them a break from treatment, leads to better results than waiting until they get sicker to start them on treatment. These long-term findings from the CHER trial, which took place in South Africa, were published in the Lancet on 9th November.

Infants with HIV are at a high risk of dying or their disease getting worse. Early treatment with antiretroviral therapy (ART) has dramatically changed this. However, treating infants with ART has its own problems: starting children on ART is a lifelong commitment, and costs money. The drugs can have side effects. The HIV may develop resistance to the drugs that are currently available, and the earlier they start treatment the more time there is for resistance to develop.

At the time the trial was started (in 2005), infants in South Africa were not given antiretroviral treatment until they were sick or their immune system showed signs of damage. The CHER (Children with HIV Early antiretroviral) trial tested starting infants on treatment early, but then stopping their treatment for a while, compared with waiting until they showed signs of harm from HIV, then continuing treatment for life. Researchers followed up these children for up to six years, to see how well they survived and remained free of disease progression.

The trial found those infants who started ART straight away for one or two years, and then stopped, had better survival by their 5th birthday than those who did not start treatment until later. Children who started ART straight away (between 6 and 12 weeks) for one or two years and then had a break were more likely to survive longer and with no symptoms than those who started treatment later. The results also suggest that children who had early ART for two years may be able to have a slightly longer break from treatment and tended to have better health outcomes than those who had early ART for only one year. One third of those who stopped ART were still well enough not to be on treatment 3 years later, and had very good immunity levels.

CHER is the first and longest randomised trial assessing practical strategies (that can be implemented in low and middle income countries) for the management of infants with HIV. The encouraging results highlight the importance of rolling out testing for HIV in infancy, so babies can be started on treatment early.

The early results from the trial have already had an impact across the world, as back in 2008 they led to a change in the World Health Organisation’s guidelines. Instead of waiting until an infant’s disease progresses to a certain level before starting treatment, treating infants immediately was shown to significantly improve survival. These new results reinforce the early results. They also provide encouragement that for some children who started ART early (and especially if they took it for 2 years), treatment can then be stopped safely for up to 3-4 years, if the child’s health and immune system is monitored well.

There was no early treatment group who then took ART continuously in the CHER trial. This raises a number of questions, including how early treatment followed by a break might compare with early continuous treatment. We also need to find out what the optimum length of initial treatment is. This trial looked at one year and two years, and two years seemed better. But it may be that treating children for 3-5 years could mean that children could have an even longer break before they need to go back on treatment for life.

The trial was carried out in Cape Town and Johannesburg in South Africa by the Perinatal HIV Research Unit of the University of Witwatersrand, the Children’s Infectious Disease Clinical Research Unit of Stellenbosch University, and MRC Clinical Trials Unit, London, as part of the CIPRA-SA programme. The research was funded by the Division of AIDS, NIAID, NIH and drugs were provided by GlaxoSmithKline.