Viral Rebound Is Significantly Delayed Among Some Who Interrupt HIV Treatment

Executive Editor

You have a patient who is sick and tired of the HIV treatment grind. After years of daily medication doses, she just wants a break. Is taking a month or two off from therapy truly such a bad thing, she asks?

You may think that the SMART study provided the beginning and end of that conversation, and that the answer is conclusively that treatment holidays are a bad idea. But if new research presented at IDWeek 2014 is any indication, the current argument for or against HIV treatment interruptions may be more nuanced.

The new research in question is a six-study pooled analysis conducted by Jonathan Li, M.D., of Harvard Medical School's Brigham and Women's Hospital (who presented the findings), and a team of HIV research luminary colleagues. For the pooled analysis, Li et al looked at six AIDS Clinical Trials Group trials that followed the traditional "analytic treatment interruption" model: Antiretroviral therapy is discontinued for 12 or more weeks, creating the possibility for long periods of uncontrolled viremia. The primary outcome of these studies is typically the length of time it takes for volunteers' viral loads to climb to a specific set point.

Such studies, because of the potential for prolonged viral replication, allow researchers to mine the data in search of a better understanding for so-called "rebound kinetics," such as the time to viral rebound (and the factors that may affect it) or degree of CD4+ cell loss while off therapy.

In this case, the pooled analysis revealed that 31% of volunteers still had a viral load below 200 copies/mL four weeks after interrupting their HIV treatment. By the eighth week, this proportion decreased to 9%, and by week 12 it was down to 6%. This sounds bleak, but the fact remained that a notable minority of volunteers still had very low levels of viral replication three months after discontinuing therapy entirely.

This raises an inevitable question: What are the characteristics that set these prolonged-rebound volunteers apart from the rest? The pooled analysis explored some of the likeliest factors, and found that people tended to spend more time with a viral load below 200 copies/mL after interrupting treatment if they:

  • Had initiated treatment during acute or early HIV infection instead of during chronic infection (13% versus 3% at 12 weeks post-interruption, P < .01).
  • Had been on a regimen containing an non-nucleoside reverse transcriptase inhibitor (NNRTI) instead of one that did not (44% versus 13% at four weeks post-interruption, P < .001; by week 12, however, this differential had vanished).

In addition to virologic assessments of how volunteers handled their off-treatment period, Li et al examined immunological effects as well. Specifically, they explored how much CD4+ cell loss occurred between the day the volunteer interrupted antiretroviral therapy and the day that person's viral load passed an established threshold (the lowest threshold they used was 200 copies/mL). They found that the lower a patient's nadir CD4+ count, the greater CD4+ loss that person experienced while off therapy and while viral load remained below 200; those with a CD4+ nadir above 500 experienced the least amount of loss, those with a nadir between 201 and 500 experienced slightly more loss, and those with a nadir below 201 experienced even more loss. (Specific values were not provided in the presentation; the data were presented as a series of bar graphs.)

This pooled analysis was not conducted to provide clinicians with practical guidance regarding which of their patients might fare better on a preplanned treatment interruption. In fact, it was grounded in a need for additional foundational research on how to measure the success of experimental HIV cures. "The success of interventions aimed at achieving [antiretroviral therapy]-free HIV remission will ultimately be judged by their ability to show clinically meaningful results in HIV treatment interruption trials," Li said during his presentation.

Even if this study was intended to inform treatment interruption decisions, it would be unable to reach many solid conclusions due to being too limited in its scope, the number of patients included in the analysis (235 in total), the number of factors explored, the variety of antiretroviral regimens used in initial therapy, the time volunteers spent on treatment before commencing their interruption, and the amount (and consistency) of baseline data available on the study volunteers.

Instead, what this research provides is a signpost for further exploration -- not only for determining the groundwork for research into HIV cure efficacy, but also for identifying factors that reduce the likelihood of immediate virologic rebound and immunologic degradation when a person interrupts therapy. Li et al plan to continue to examine the former; the appetite for renewed research on the latter is less certain.

[CORRECTION 10/24: The original version of this article inaccurately identified the study by Li et al as a meta-analysis. It has been corrected to refer to the study as a pooled analysis.]

Myles Helfand is the editorial director of and

Follow Myles on Twitter: @MylesatTheBody.