Controversy regarding strategic treatment interruptions in young children who start ART during infancy

This article is more than 10 years old. Click here for more recent articles on this topic

Two studies generated conflicting data and rather mixed messages about whether it is safe or advisable to strategically interrupt treatment in children (STIC) who have initiated antiretroviral therapy (ART) during infancy – in order to reduce the challenges of adherence and the risks of long-term toxicity and resistance associated with lifelong therapy – at the 19th Conference on Retroviruses and Opportunistic Infections (CROI) in Seattle.

“Treatment interruption was not feasible for our study population,” said Dr Dalton Wamalwa of the University of Kenya who presented the results of the Optimizing Pediatric HIV-1 Therapy 03 (OPH03) study.1 In contrast Dr Mark Cotton of Stellenbosch University in South Africa, who presented final results from the pivotal CHER study (see below), reported that, “stopping treatment at the first and second birthday appears safe.”2

But there were some caveats. Dr Cotton noted there was a trend towards better results in those remaining on treatment until at least their second birthday (and that there had been no comparison to longer or continuous treatment in the study). Furthermore, the time that the children could remain safely off-treatment before re-starting was not very great (a median of 33 weeks or 70 weeks depending on the arm), which led Dr Andrew Prendergast of Queen Mary College, University of London, to ask whether such a duration of treatment interruption is long enough to be worthwhile, during a presentation he gave on Wednesday on the controversies of early ART initiation in children.3


treatment interruption

Taking a planned break from HIV treatment, sometimes known as a ‘drugs holiday’. As this has been shown to lead to worse outcomes, treatment interruptions are not recommended. 

CD4 cell percentage

The CD4 cell percentage measures the proportion of all white blood cells that are CD4 cells.

disease progression

The worsening of a disease.



first-line therapy

The regimen used when starting treatment for the first time.

While the presenters of both STIC studies stressed that their research was not the final word on the subject – and that further analysis and research is needed – Dr Prendergast’s presentation was followed by some heated discussion on the ethics of treatment interruption in children.

Challenges of ART in infants and young children

But that such a discussion could be had is testament to how far the management of children living with HIV in resource-limited settings has come.

HIV infection in infants and children poses several major challenges, as Dr Prendergast pointed out in his presentation, First, HIV-infected infants have very high viral loads and, most importantly, very rapid disease progression – and there are no reliable markers to distinguish between which infant will be a fast or slow progressor. They also have fewer treatment options due to lack of approved paediatric formulations, and many children have prior NNRTI-exposure (nevirapine) due to its use to prevent vertical transmission. Second- and third-line regimens are more difficult to construct as a consequence. Finally, adherence can be quite challenging in children, and there are also questions regarding the costs and toxicity of treatment.

Fortunately, a handful of randomised controlled studies have now shed some light on how to respond to clinical management challenges in children living with HIV, perhaps most importantly, when to start treatment.

The answer, according the Children with HIV Early Antiretroviral Therapy (CHER) trial study is as soon as you possibly can.

In 2007, after a median of only 32 weeks of follow-up, the trial’s Data Safety and Monitoring Board recommended a halt to enrolment in the deferred treatment ARM after the study found that treating asymptomatic HIV-infected infants with antiretroviral therapy (ART) within the first six to 12 weeks of life – rather than waiting until they show signs of immunological or clinical deterioration – dramatically reduced the risk of early death by 74%.4

This led to a change of guidelines in all settings to recommend treatment in the first two years of life and a new emphasis on early infant diagnosis.

Given the benefits of early ART, why would anyone consider strategically stopping treatment in children living with HIV?

However, “lifelong ART is really not a realistic goal in terms of toxicity, resistance and cost,” said Dr Prendergast. Consequently, a number of trials, including the other CHER study, which is ongoing with some modifications, are investigating the feasibility of different structured treatment interruptions in children who started treatment in infancy.

Many people in the paediatric world, particularly the US, see the approach as questionable.

Much of the concern arises from data from studies describing unplanned treatment interruptions, which have shown that stopping treatment in children may cause a rapid increase in viral load, a decrease in CD4 cell count that is frequently persistent, and an increased risk of clinical disease progression.5

However, there are a few reasons why STICs might be different in children in resource-limited settings. One is that early ART in infancy may preserve or salvage thymic function, and lead to different responses to treatment interruptions than in adults, as Dr Wamalwa mentioned in his presentation — in fact, it has been postulated that unlike in adults, treatment interruption in children might stimulate a CD8 cell response that may contribute to virologic control.

Two studies were presented at CROI providing further evidence on the question of treatment interruption after early antiretroviral treatment.


OPH03 was an unblinded randomized study that compared growth and serious adverse events among children who started empiric ART in infancy (below 13 months of age) and on treatment for at least two years. Children were randomized to treatment interruption vs continuing ART if their CD4 >25% and their growth had normalised while on treatment.6 The children were to be restarted on ART if their CD4% fell below ≤25% (as per WHO guidelines) or dropped more than one third below peak.

Of 121 children on ART in the study, 75 had more than 24 month of treatment, and 42 were eligible for randomization (21 per arm). Very soon, however, the DSMB recommended stopping the study because of the high proportion (86% or 18/21) who restarted ART, 16 of whom (76%) met CD4 criteria – the majority (14) within just three months of going off treatment.

The treatment interruption durations were probably too short to show whether there were differences in growth or serious adverse events (only one SAE was observed in each arm). The differences were simply immunological and virological. At the 3 month time point after STI, CD4 dropped markedly while viral load rebounded quickly before stabilising.

A faster restart was significantly associated (p = 0.04) with lower CD4 percentage at the time of randomisation (though not pre-ART) – having a higher CD4 percentage at planned treatment interruption correlated with a smaller absolute decline in CD4 percentage post-treatment interruption.

Dr Wamalwa, presenting the study, noted that WHO guidelines mandate that children should be on ART if their CD4 percentage falls below 25%, so using a lower restart criteria, such as in PENTA where ART was restarted if CD4% fell below 20%, was simply not possible in this population. Indeed, more regular monitoring in European settings might make it less risky for a child to have a somewhat lower CD4 percentage, but given the rapid rate of progression in young children in Africa, and infrequent monitoring, it may not be a safe strategy in Kenya.

Structured treatment interruptions later in childhood might also be possible, Dr Wamalwa suggested, and “infants detected while asymptomatic or with higher CD4 percentage may also be a better group for treatment interruptions.”


Planning for the CHER study began much earlier, and the investigators clearly expected that scarcity of ART, costs and a ‘commitment to lifelong ART’ might prove a stumbling block to the widespread acceptance of any findings the study might produce about early treatment. Fortunately that has not been the case, and early ART in infants has become the standard of care.

Nevertheless, with some modifications, the CHER study continued, in order to compare structured treatment interruptions, which were planned to take place after the infants reach either their first or second birthday.

CHER Trial: Part A n = 375

HIV-infection diagnosed before 12 weeks and CD4% > 25%


Defer ART until clinical progression or CD4% drop



Early ART to 40 weeks; then STOP, until progression



Early ART to 96 weeks; then STOP, until progression


1st-line ART: Kaletra + ZDV + 3TC

ART (start or re-start) when CD4% < 20% or clinical event

Follow: Up to 6 years

Primary endpoint: Time to failure of 1st-line ART

ART initiation (in the deferred arm while it lasted) and re-initiation (at which point the child began continuous ART) were based on WHO immunologic criteria (CD4% <25%; <20% after infancy) or CDC severe B or C clinical disease. The composite primary endpoint was time to failure of first-line ART (clinical progression or toxicity or a fall in CD4% below 20% initially, changed to <25% after WHO guidelines changed) or death. Secondary endpoints, included:  the cumulative rate of disease progression and hospitalization, grade 3 & 4 adverse events, development of ART resistance. Once the study was modified an extra 34 infants enrolled (ART-40W and ART-96W).

At baseline, the population was well matched across arms, most of the infants were around seven weeks of age, there was a slightly higher proportion of female infants, and the median CD4 percentage was 34-36%.

The median follow-up was 249 weeks (4.8 years), and the maximum was 309 weeks (5.9 years). As already noted, for the primary outcome, death or failure of first-line ART, there was a significant reduction for those who had immediate therapy (the ART-96W and combined arms) versus the deferred arm(although the reduction was not as great since clinical events were reduced in frequency once people in the deferred arm started treatment).

                                        Death or failure of first-line therapy

Study arm

HR relative to ART deferred

95% CI, p value

ART until 40 weeks


(0.46-1.17, p = 0.19)

ART until 96 weeks


(0.35-0.96, p = 0.03)

Combined treatment interruption arms


(0.43-0.98, p = 0.04)

                              Progression to severe CDC B or CDC C, or death


ART until 40 weeks


(0.3-0.8, p = 0.005)

ART until 96 weeks


(0.3-0.7, p = 0.0003)

Immediate treatment cut mortality by half and significantly reduced disease progression compared to deferred treatment. 

When the two immediate arms were compared to each other, there was no significant difference between them, although there was a trend to fewer clinical events on the ART-96W arm. Most of the events occurred before the treatment interruption in ART-96W.

Clinical Events at each study step:


39 events

(34 children)

N = 143


31 events

(28 children)

N = 143

Primary ART (Over year 1)

  • Deaths – 9
  • Clinical – 6

Primary ART (over Year 1 + 2)

  • Deaths – 10
  • Clinical – 18

Interruption (Y2 +)

  • Deaths – 1
  • Clinical – 9

Interruption (Y3+)

  • Deaths – 2
  • Clinical – 0

After ART restart

  • Deaths – 1
  • Clinical – 13

After ART restart

  • Deaths – 0
  • Clinical – 1

The treatment interruptions did appear to go better in this study than in Kenya. Time to starting continuous ART (after interruption) adjusted for length of primary ART) was a median of 33 weeks (26-45 IQR) on ART-40W, and 70 weeks (35-109 IQR) on ART-96W, a difference of 37 weeks (-11-85 95% CI, p = 0.13).

“The time to restarting continuous ART… it is really quite strikingly early,” opined Dr Prendergast – even median duration off treatment on the ART-arm was longer than any of these studies have reported thus far (and it should be noted that a quarter of the week 40 and a third of the ART-96W remain off-treatment).

However, it must be remembered that – at most, these are two-year old children when the stop treatment, and they must begin treatment again before the age of four. These may not be the most critical years where there might be significant problems with adherence.

As Dr Cotton noted, ART interruption appeared safe, but so was treatment, relatively, suggesting that, at best, the strategy was cost saving compared to deferred and then continuous treatment.

Moreover, Dr Cotton stressed that further analyses are needed of virological suppression and resistance off treatment and once it is restarted. Likewise, no data were presented on the immunological response to restarting ART after interruption – and it could well be that continuous treatment could have preserved or sustained better CD4 cell percentage or CD4 cell count – and that the treatment interruption may have increased the risk of resistance. The study had no comparison to continuous ART, making it impossible to conclude whether a treatment interruption is actually the best strategic choice for a child (or just a way for programmes to save money).

Ultimately, Dr Cotton concluded, more research and analyses are needed. “Could early primary ART for longer (e.g. for 3-5 years) be better? And any future trial should compare early primary ART with early continuous ART,” he said.


Dr Prendergast reiterated these same points in his talk.

“What duration of treatment interruption would be considered worthwhile? Is it suitable for all? Certainly in the Kenyan study, those who had lower CD4 percentage at randomisation were more likely to need an early restart to therapy? Do we need to decide which children would benefit from this sort of strategy perhaps by using CD4 count? Would it be better to wait to three to five years of age before interrupting treatment, when disease progression is less rapid – maybe then the children would last longer off treatment. What is needed now is a study of immediate treatment randomised to a treatment interruption or continuous treatment,” he said.

However, this did not settle well with everyone in the audience.

“I’m shocked, shocked,” said Dr Charlie Van Der Horst, of the University of North Carolina. “You don’t stop treatment. In adults we learned a long time ago that treatment interruption is a bad thing. It is completely unethical to stop treatment on an infected child. This is a virus affecting developing brains. It’s totally crazy. If I were on a DSMB or IRB, I would never approve such a study – it has been proved in adults many years ago. Stopping therapy is Middle Ages.”

Dr Prendergast could only respond that continuous treatment in infants who start treatment at 6 weeks of age is qualitatively a different thing when one thinks that they will have to remain on it into their thirties, into their middle age, and old age.

Treatment interruption studies in children are likely to remain controversial and closely scrutinised as further results from these studies are presented.

  1. Wamalwa D et al. Treatment Interruption in Infants following 24 Months of Empiric ART: Kenya. 19th Conference on Retroviruses and Opportunistic Infections, Seattle, abstract 27, 2012. This abstract is available on the official conference website.
  2. Cotton M et al. Early ART followed by Interruption Is Safe and Is Associated with Better Outcomes than Deferred ART in HIV+ Infants: Final Results from the 6-Year Randomized CHER Trial, South Africa. 19th Conference on Retroviruses and Opportunistic Infections, Seattle, abstract 28LB, 2012. This abstract is available on the official conference website.
  3. Prendergast A et al. ART for the HIV+ Infant: Controversies and Consequences of Early Initiation. 19th Conference on Retroviruses and Opportunistic Infections, Seattle, abstract 112, 2012. This abstract is available on the official conference website.
  4. Violari A et al. Early antiretroviral therapy and mortality among HIV-infected infants. N Engl J Med. 359(21):2233-2244, 2008.
  5. Gibb DM et al. Immunologic changes during unplanned treatment interruptions of highly active antiretroviral therapy in children with human immunodeficiency virus type 1 infection. Pediatr Infect Dis J. 23(5):446-450, 2004.
  6. Wamalwa, op.cit

Webcasts of the conference sessions, Critical treatment issues in women and children and Recognizing risk in HIV-exposed and -infected infants and children, are available through the official conference website.