Analytical treatment interruptions: minimising the risks and maximising their scientific value

Recommendations for the safe and appropriate use of ‘analytical treatment interruptions’ have been developed by a group of HIV experts, including clinicians, patient advocates, bioethicists, social scientists and representatives from regulatory authorities, and recently published in The Lancet HIV.


For all people living with HIV, antiretroviral therapy (ART) unquestionably represents the best opportunity to achieve viral suppression, conserve or recover immune function, and prevent HIV-related diseases. ART should therefore be taken indefinitely.

However, ART interruptions may be necessary in certain challenging cases: treating another illness, unbearable ART side-effects, etc. Also, research that tests promising strategies like immune-based therapies and aims to achieve HIV eradication or ART-free control of HIV viral load may require that HIV-positive volunteers interrupt their ART and be closely monitored: an approach referred to as an ‘analytical treatment interruption’, or ATI.

In these studies, the use of ATIs is justified because, so far, no biomarker (a sign detected in the blood, for example) can predict post-treatment viral control and, consequently, help assess the efficacy of the tested strategies. Until such a biomarker is discovered, researchers can only rely on how long the HIV viral load remains suppressed, from the moment ART is interrupted to the first viral load rebound, as a study endpoint.


analytical treatment interruption (ATI)

As part of a research study, when study participants are requested to interrupt their antiretroviral therapy (ART) and be closely monitored. Most of these studies are in the field that aims to eventually achieve ‘ART-free remission’ or ‘HIV cure’. They usually look at the time it takes for the viral load to rebound after the interruption, following which ART is restarted.


The presence of virus in the blood.



In a clinical trial, a clearly defined outcome which is used to evaluate whether a treatment is working or not. Trials usually have a single primary endpoint (e.g. having an undetectable viral load) as well as a few secondary endpoints, covering other aspects of treatment safety, tolerability and efficacy.


The ‘HIV reservoir’ is a group of cells that are infected with HIV but have not produced new HIV (latent stage of infection) for many months or years. Latent HIV reservoirs are established during the earliest stage of HIV infection. Although antiretroviral therapy can reduce the level of HIV in the blood to an undetectable level, latent reservoirs of HIV continue to survive (a phenomenon called residual inflammation). Latently infected cells may be reawakened to begin actively reproducing HIV virions if antiretroviral therapy is stopped. 

exclusion criteria

Defines who cannot take part in a research study. Eligibility criteria may include disease type and stage, other medical conditions, previous treatment history, age, and gender. For example, many trials exclude women who are pregnant, to avoid any possible danger to a baby, or people who are taking a drug that might interact with the treatment being studied.

However, viral rebounds and potential consequences raise concerns and call for caution. Potential consequences can include the development of a resistant virus, an acute retroviral syndrome (signs of HIV infection, reminiscent of the primary stages of the infection), thrombocytopaenia (low platelet counts), AIDS-defining events and HIV-related events.

Another issue faced by the scientific community is that while research employing ATIs is expanding worldwide, clinical trial designs are so heterogeneous that comparing their results can be challenging.

A group of HIV experts gathered in July 2018 in the US to discuss these concerns and reach consensus on recommendations for the use of ATIs in research. Their recommendations are mostly based on expert opinion, but are likely to have a significant influence on future clinical trials.

Separately, in November 2018, community recommendations on research involving ATIs were published.

Informed consent and scientific rationale

One of the first principles is that investigators must ensure that an ATI and its potential benefits and risks are adequately understood by participants. Participants must receive adequate information.

As an example of a potential risk to explain, the HIV reservoir might increase in size. The reservoir is a group of HIV-infected cells that have not produced new HIV (latent stage) for many months or years, but could start doing so when triggered by an event such as an infection-induced inflammation, if HIV is not controlled by ART.

Nonetheless, the experts’ review of the available evidence suggests that this specific risk is limited. One study had shown that integrated HIV-1 DNA (the virus’ genetic information integrated in the cell’s) remained elevated at least for two years after ART reinitiation. However, in another study a significant HIV reservoir expansion occurred only after 60 weeks of ATI; and there was a return to the previous reservoir levels within six months of restarting the same ART that had been taken prior to ATI.

Separately, the experts observed that the risk of viral mutations increasing during an ATI (and of resistance to ART) was also limited. 

Another principle to emerge from the experts’ discussions was that it is up to an ATI trial investigator to show that there is a strong scientific rationale for the trial. The investigator must determine pre-defined go/no-go criteria (to help decide if the strategies should continue to be tested or not).

Also, ATIs should not be used to only generate hypotheses, in the absence of supporting data. In other words, no therapeutic approach should be tested in an ATI clinical trial without available animal/pre-clinical data suggesting a potential benefit of the approach.  

Who can take part?

Mitigating risk for potential trial volunteers was a constant concern of the experts, although they did not want ATI studies to have inclusion criteria so stringent that few people living with HIV could take part. These considerations brought the group to agree on a “description” of what type of volunteers should be enrolled in the ATI studies:

  • Healthy individuals with well-controlled HIV who do not have serious co-morbidities. They noted that people with ART-controlled HIV infection, no symptoms and many treatment options are more similar to HIV-negative persons than to people with life-threatening issues, in terms of the risks considered to be “acceptable” in a research study.
  • Individuals with a functional immune system, who can tolerate a period of viraemia (even of high viraemia), usually with stable CD4 cell counts of 500 cells/mm3 or above. In studies where the pre-clinical data indicates the intervention is of a lower risk, a CD4 count threshold of 350 cells/mm3 or above may be appropriate.

As the influence of sex, gender, race, ethnicity and geographic location on ATI outcomes requires further investigation, the experts underlined how important it is to develop strategies to recruit participants with a range of demographic characteristics.

As regards children with HIV, the group’s opinion was that, although they might benefit the most from ART-free approaches, their unique complexities (stages of immune and neurologic system development, potential influence of vaccinations, dependency on adults, etc.) called for separate, dedicated age group-specific recommendations.

Who should not take part?

Exclusion criteria were carefully addressed. Generally speaking, active co-infections, cancer, neurological concerns, current ART resistance, cardiovascular disease, a history of AIDS-defining illness, CD4 nadir (lowest level ever reached) below 200 cells/mm3, pregnancy, renal or liver disease, risk of HIV transmission to sexual partners (see below) and children under two years of age, fell in this category.

Importantly, some situations may allow for flexibility:

  • Individuals who have been cured of hepatitis C or who have cleared the virus naturally do not need to be excluded.
  • While active tuberculosis is a strict exclusion criterion, the latent form is not. However, caution is advised, as the disease can reactivate (one option could be to treat latent TB when detected).
  • A history of some types of cancer (colon, prostate, breast, even systemic cancers such as lymphomas in very specific cases) might not result in exclusion, depending on clinicians’ assessment of the risk-benefit ratio.

Special attention was paid to the risk of HIV transmission during an ATI (it is known that HIV infected cells can re-populate genital compartments). Consensus was reached that participants in ATI studies should receive clear and comprehensive counselling on transmission risks. This counselling should include information on pre-exposure prophylaxis (PrEP), post-exposure prophylaxis (PEP) and HIV testing, to give to their HIV-negative sexual partners.

This may reduce the risk of HIV transmission and also the risk of legal charges following transmission to a sexual partner. But because of their ethical obligations to protect participants’ confidentiality, the experts suggested that ATI study investigators would not deliver PrEP or PEP to their partners.

How should participants be monitored?

ATI study participants must be closely monitored: weekly viral load measurements for 12 weeks, then every other week, but more often if necessary (for example, once the viral load becomes detectable). The rationale for a more intense monitoring during the first 12 weeks is that so far, most viral rebounds in ATI studies have occurred within this timeframe.

While viraemia is the key measurement, CD4 counts should also be measured every two weeks. Clinical symptoms should be checked throughout the trial, especially following viral rebound, due to the possibility of acute retroviral syndrome.

The experts also agreed that individuals who started ART in the very early phase of their HIV infection and may not have developed detectable anti-HIV antibodies (“seroconverted”) would require specific attention, as they are at greater risk of acute retroviral syndrome.

The writing group also felt that it was crucial to formally collect information on the psychosocial and lived experience of study participants during ATIs. In their view, researchers should assess patients’ motivations, needs, concerns, and perceptions of benefits and risks (including fear of transmitting the virus, anxiety about not being on treatment, etc.). They should also consider participants’ tolerance for longer ATIs, as research might well be moving in that direction.

When should treatment be taken again?

The best time to re-initiate ART was of course a key point in the discussions. The general conclusions were that ART should be restarted if:

  • a participant or their HIV clinician wants to, regardless of the reasons
  • a participant becomes pregnant
  • ART is considered necessary for non-HIV related causes (for example, cardiovascular health)
  • HIV-related symptoms appear.

Quite obviously, viraemia is an important ART restart criterion. But the choice of virologic endpoint should depend on the study objectives, rather than the same endpoint being used in all studies. For example, although time to viral rebound might be the safest endpoint (preventing complications of an ATI), restarting ART as soon as viral load reaches 100,000 copies/ml might not allow for thorough explorations of the immune mechanisms triggered by a tested strategy, as they may only occur in the presence of a high viraemia during several months.

An alternative criterion for restart might be the duration of viraemia, or its peak level. The experts did not define any universal values, but felt that in general, duration might be more important than peak.

CD4 concentrations can also be envisaged as ART restart criteria, but not in competition with viraemia. A CD4 count below 350 cells/mm3 or CD4 percentage of less than 15% were proposed. Alternatively, the percentage of decline from the starting CD4 levels (for example, 30-50% decline) could be considered. However, whatever level is chosen, it should be sufficiently different from participants’ CD4 concentrations when they entered the study.


Overall, these recommendations indicate that decisions about ATIs may differ from one study to another, and even from one individual participant’s case to another, therefore implying a need for a certain flexibility with respect to inclusion and exclusion criteria. This underlines the need to give participants adequate information at all stages of the studies.

Among other information tools, the informed consent document can play a major role in this process, but only if the researchers commit to support it, by double-checking that participants have read it, been given time to ask questions, etc., particularly in the early phases of ATI trials. As well as warning about potential physical ATI risks, the informed consent document should address the social, financial and psychological risks (having detectable viral load, fear of transmitting HIV, etc.). Investigators must be careful not to encourage unrealistic optimism in participants, in terms of what a trial may achieve.

The outcomes that researchers hope they may eventually achieve have for many years been oversimplified and described with the words “HIV cure”. But as most current ATI studies are not targeting such outcome, the term is misleading. So, what to use? Remission? Drug or ART-free long-term virologic control? Durable viral load suppression? The experts decided that these suggestions, and probably many others, should be considered for review by the community of people living with HIV.   


Julg B et al. Recommendations for analytical antiretroviral treatment interruptions in HIV research trials—report of a consensus meeting. Lancet HIV 6:PE259-E268, 2019. (Full text freely available).