The strategy of structured treatment interruption (STI) has become quite popular in the field of antiretroviral therapy over the past year. Functionally, it can be construed as having three goals:

1. To stimulate the host's immune responses to HIV by allowing limited exposure to endogenous strains of the virus
2. To repopulate the viral swarm with wild–type virus, which may be more responsive to antiretroviral therapy
3. To reduce the long–term toxicity of antiretroviral drugs by limiting cumulative exposure to them.

Considerable enthusiasm has been generated by the data of Rosenberg, Walker and colleagues of Harvard University, which was published last year in Nature [1]. In brief, they established a protocol for the treatment of acute HIV infection, wherein individuals were diagnosed and treated prior to the development of a complete Western blot antibody pattern. Antiretroviral therapy was discontinued if the plasma viral load was below 50 copies/mL for an arbitrary duration of 8 months. Therapy would be re–initiated if the viral load increased above 5,000 copies/mL for three consecutive weeks, or above 50,000 copies/mL at a single time. Four of 14 patients treated in this study were able to stay off treatment indefinitely after the first STI, with 3 additional patients achieving this threshold after a second or third STI. It is thought that, if this approach is to work, it will be easiest to demonstrate in acute HIV infection, where the viral swarm is narrow, viral tissue reservoirs are least well established, and the immune system is relatively intact.

These data were largely supported by observations in 10 European patients in whom treatment was initiated during acute HIV infection and continued for 7 to 35 months (Abstract 290) [2]. In follow–up, 4 have maintained plasma viral loads below 50 copies/mL off therapy for 18 to 40 months. A similar approach has been applied to a cohort of 14 patients in New York City (Abstracts 288 and 359) [3, 4]. All were treated during acute HIV infection, and responded dramatically to HAART. On therapy, the median CD4 cell count was 811 cells/µL, and most patients had robust anti–HIV immune responses, as measured by increased CTL activity (10/14) or CD4 proliferative responses (7/14). Discouragingly, off therapy, all 14 had a virologic rebound within 6–107 days, with only 2 patients maintaining a plasma viral load below 1,000 copies/mL.

In chronic infection, a study of 12 patients has shown that sequential interruptions of therapy (for 30–90 days, based on the time to virologic rebound above 3,000 copies/mL) showed lower levels of rebound after each interruption (Abstract 291) [5]. A substantial increase in HIV–specific CD8 cell frequencies was observed in 5/12 patients. In a controlled study of HAART with STI (n = 10, 3 cycles over one year) vs. a control group in which HAART was simply stopped (n = 20), the patients undergoing STI had a 0.78 log₁₀ lower plasma viral load than the controls at the end of the follow–up period. In a Swiss–Spanish study of 128 patients (Abstract 357), a strategy of "2 weeks off/8 weeks on" was employed [6]. After 4 cycles, 9/54 (17%) patients maintained virologic suppression after a further 12 weeks off therapy, the majority of whom had initiated therapy during the earliest stages of their infection.

The STI strategy is not without risk, however. In a group of 19 patients on a PI–based regimen, STI for a mean period of 18 weeks was accompanied by a decrease in CD4 cell counts of 95 cells/µL, and an increase in plasma viral load of 0.74 log₁₀ copies/mL (Abstract 292) [7]. Other data suggest that the CD4 cell count decline off therapy could be approximately 16 cells/month (Abstract 355) [8], with a doubling time of HIV in the plasma of 1.6 days (Abstract 360) [9]. The majority of patients (18/19) had a predominant circulating isolate that was sensitive to PIs. However, when therapy was re–started, only 9 patients exhibited a virologic response, mainly related to the fact that they were naive to NNRTIs, a fact that has little or nothing to do with the STI having occurred. Other patients rapidly showed the emergence of PI–resistant isolates, which are likely to have been still present as minority isolates during the STI. This hypothesis was confirmed in another elegant study in which cloning experiments quantified this phenomenon (Abstract 293) [10].

The optimal duration of a proper STI has not been determined. It is quite clear that there is great interaction variability, especially in the setting of chronic infection. In one study of 68 patients responding to therapy, Kaplan–Meier estimates of the median time for the CD4 cell count to decline back to pre–treatment levels was 24.6 weeks, with great inter–patient variability (Abstract 361) [11]. This is probably too long an interval to wait to re–initiate treatment. In data pooled from 6 clinical cohorts (n = 252), a reduction in CD4 cell count of 180 cells/µL was observed after a median treatment interruption of 4.3 months (Abstract 365), suggesting that this interval is also too long, but the specific interval that would maximize the potential benefit of STI while minimizing the risk has yet to be determined [12]. A randomized controlled trial has been undertaken to compare intermittent and continuous HAART in patients with maximal virologic suppression on therapy (Abstract 364) [13]. The STI protocol consists of "4 weeks off/2 weeks on" HAART for 22 months. In preliminary data after 2–5 cycles, 13 of 14 STI patients have detectable plasma viral loads at each interruption, and 13 of 14 achieved maximal suppression when drug therapy was re–initiated (only one persistent breakthrough). The majority of patients (11/14) experienced lower levels of plasma viremia with repeated STI cycles, 7 with decreases of 0.5 log₁₀ copies/mL or greater. A 17% decline in CD4 cell count was measured at the time of the first treatment interruption, but this was reduced to 3–9% after subsequent cycles. These preliminary results suggest that this trial (with an appropriate parallel control group of patients who are not undergoing an STI) will go a long way towards helping us understand if the STI concept will have any clinical validity. Until these results are known, the STI strategy remains in the realm of clinical research, rather than clinical science.

In an interesting twist to the concept of intermittent therapy, 12 individuals with prolonged responses to therapy were assigned to receive cycles of d4T/3TC/IDV/RTV for 7 days, followed by 7 days off drug for 24 months (Abstract 354) [14]. Preliminary data suggest that virologic suppression and immunologic response to therapy was maintained for the entire period of observation. If this approach is validated, it could represent a way to reduce drug exposure (and drug cost) without compromising efficacy. A larger trial is clearly needed.

In conclusion, the concept of STI is interesting, but may have limited long–term benefit outside of acute/early HIV infection. In more established infection, its use must remain limited to clinical trials, which much be properly designed to help address the key issues in the field, such as:

1. Appropriate duration of therapy prior to the STI
2. Criteria for re–initiating therapy after an STI
3. Optimal regimen to use before and after the STI
4. Predictors of success of the strategy in individual patients
5. Immunologic correlates of the approach

Hopefully, when we review the CROI next year, we will have improved our insight into STIs, and developed a better understanding of their ultimate clinical utility.

References