Intermittent versus continuous HAART

Paul E. Sax, M.D. for

A strategy of interrupted therapy – dubbed “structured intermittent therapy” – gained a major imprint of respectability at this past summer’s International AIDS Conference in Durban when Anthony Fauci cited his group’s preliminary work in this area.

Numerous other terms have been applied to this practice, including “structured treatment interruption,” “strategic treatment interruption,” and “supervised intermittent therapy”, but the basic principal underlying all of them is that antiretroviral therapy need not be continuous. In these two posters, Mark Dybul from Tony Fauci’s lab presented their ongoing work in this area.

In the first poster [1], the question posed was, “Can 50% less antiretroviral therapy given as short-cycle Ôstructured intermittent therapy’ maintain suppression of viral load to reduce cost and toxicity of treatment?” Patients were eligible to join the study if they had CD4 cell counts >300 and viral load <500 for more than six months, and had a viral load of <50 copies at the time of screening. The treatment strategy planned was seven days on/seven days off treatment for 24 months or until treatment failure (viral load >500 twice or CD4 decline by 25% or more) ensued.

At entry, all 10 patients were receiving the same antiretroviral therapy, which was d4T/3TC/ritonavir/indinavir. At the end of 3-11 months of follow-up, 8/10 maintained a viral load <50 during the course of study. One patient had a brief virologic rebound when he came off treatment for 10 days; a second patient failed to resume treatment after three weeks, and the viral load increased to 20,142 copies/mL. Both patients achieved virologic suppression when they resumed treatment. In the entire group, CD4 and CD8 counts did not change.

An analysis of viral reservoirs checking for amount of proviral DNA and replication competent HIV both pre and post starting the intermittent treatment cycles showed no apparent increase in the reservoir. Furthermore, no decreased susceptibility to antiretroviral agents was detected when cultured virus underwent resistance testing.

Based on these preliminary data, the investigators concluded that it may be possible to safely decrease the amount of antiretroviral therapy administered by 50% without untoward effects on virologic control or immune status. Such a strategy would be particularly appealing in resource-poor settings or if it were shown to be associated with lower cumulative drug toxicity. Given the small size of the sample presented here, however, this particular approach awaits further follow-up on a larger number of patients.

The group’s second paper [2] employed a different study design: taking patients on suppressive antiretroviral therapy and randomising half to continue that treatment, and half to begin a schedule of intermittent treatment consisting of one month off therapy followed by two months on. To be eligible, patients had to have a CD4 cell count >300 and a viral load <500 for greater than three months, as well as a viral load <50 at the time of study entry. The primary endpoint was the time to viral load rebound or CD4 decline, with secondary endpoints CD4 and CD8 t-cell mediated HIV-specific immune responses.

Although a total of 70 patients will ultimately enrol, preliminary results on the 47 thus far participating were presented. At baseline, the patients randomised to intermittent versus continuous therapy were similar regarding the mean CD4 nadir, CD4 count at study entry, and peak pre-treatment viral load. For the fifteen patients in the intermittent therapy group for whom three to six cycles of interruption had been completed, four different patterns were observed:

3/15: each treatment interruption was associated with a 0.5 log or greater reduction in peak viral load

2/15: each interruption had a .5 log or greater increase in viral load

5/15: interruptions had variable viral load rebounds

4/15: rebounds were similar each time

Patients who were in the SIT group had a similar number of measurements >50 copies/mL (“blips”) during treatment as did the continuous therapy group. CD4 cell counts did decline during the treatment interruptions, but these rebounded to baseline during resumption of therapy. Of some concern, one (out of five) patients on efavirenz who was in the intermittent treatment group developed genotypic and phenotypic resistance to this agent.

Based on these preliminary results, the investigators concluded that there was no clear pattern of virologic rebound in these chronically infected patients who undergo this particular schedule of treatment interruption. While the stable CD4 cell counts overall was a promising sign, even the one patient who developed NNRTI resistance during the study raises concern about the potential for resistance with intermittent therapy. This may be particularly problematic for drugs with long elimination phases, such as efavirenz. The bottom line from these two studies of intermittent treatment is that STI remains a strategy under investigation, and cannot yet be endorsed for routine clinical practice.



  1. Dybul M, Chun TW, Yoder C et al. Short-Cycle Intermittent HAART: A Pilot Study. 8th Conference on Retroviruses and Opportunistic Infections; February 4-8, 2001; Chicago, Illinois. Abstract 354.
  2. M. Dybul, C. Yoder, M. Belson et al. A Randomised Controlled Trial of Intermittent Versus Continuous HAART. 8th Conference on Retroviruses and Opportunistic Infections; February 4-8, 2001; Chicago, Illinois. Abstract 364.

Links to other websites are current at date of posting but not maintained.