Structured Treatment Interruptions Workshop Summary
January 31, 2001
Martin Markowitz, MD
Aaron Diamond AIDS Research Center, New York, NY
A prospective study observed 8 patients who started treatment when newly infected (within 54 days of symptoms) and had been on HAART with suppressed VL (no more than one blip per year) for about 3 years when treatment was interrupted. Four patients had received ALVAC vaccination.
The pre-treatment median HIV RNA was 5.0 log copies/mL and CD4+ cell count was 498/mm3. At interruption, all patients had HIV RNA <50 copies/mL and the mean CD4+ count was 823 cell/mm3.
After interruption, all experienced viral rebound, one patient with a peak VL of 4.3 log copies/mL.
In five of eight patients, genetic characteristics of the rebound virus were identical to virus present during the individual's primary infection, and to virus isolated from latent reservoirs during treatment. Genetic characteristics of rebounding virus from three other individuals diverged from virus isolated during primary infection, and from latent reservoirs, but corresponded with minor viral variants detected in lymphoid tissue while on treatment.
Even after prolonged periods of near or complete suppression, rebound virus can arise from either archives of the founder strain in latent reservoirs or from inadequately suppressed strains that evolved in protected compartments in lymphoid tissue.
Augmentation of HIV-1-specific Immune Responses after STI
HIV-1 specific cytotoxic T lymphocytes (CTL) respond weakly to a narrow range of epitopes during primary infection. After treatment with HAART, the frequencies of CTL increase but the strength of response remains weak. The breadth of epitope recognition remained narrow in individuals treated before and after seroconversion, although a broader response was observed in individuals treated at the time chronic infection was established. Individuals treated during primary infection preserved HIV-1-specific CD4+ T cell responses as well as a more homogenous viral population compared to those who were treated later.
Treatment interruptions were performed to determine if HIV-1 specific immune responses could be enhanced to levels sufficient to control viremia. Five individuals treated during primary infection who had suppressed viral load <50 copies/mL for over eight months with no evidence of drug resistance were offered interruption. Treatment was restarted if viral load exceeded 5000 copies/mL on three weekly determinations or reached 50,000 copies/mL on one determination.
All five patients experienced rebound and restarted treatment during the first interruption of treatment. After a second interruption, all patients spontaneously controlled viremia for as long as four months. HIV-1-specific CTL responses were stronger, with new epitopes recognized during the interruptions. These responses were preserved after restarting treatment. All viral load setpoints were well below the median observed from the MACS cohort.
STI resulted in viral rebound and augmentation of HIV-1 specific immune responses that appeared to improve control of viral load during subsequent interruptions.
Cf. recent papers in Nature and J. Exp. Med.
SSITT (Spanish/Swiss Intermittent Therapy Trial)
A large prospective observational study of STI enrolled 128 patients receiving HAART who had viral load <50 copies/mL for more than six months. Four two-week interruptions were scheduled at weeks 0, 10, 20 and 30. At week 40, treatment is discontinued for all patients; the primary endpoint is viral load at 52 weeks. Eighty patients have completed four interruptions; 73 have reached 52 weeks.
Median pre-therapy baseline CD4+ count was 388 cells/mm3; HIV RNA was 4.5 log copies/mL. Four patients had a VL < 5000 pre-HAART. The median duration of therapy at entry was 25.5 months with no changes due to virologic failure. Median CD4+ count at week 0 was 727 cells/mm3; viral load <50 copies/mL.
Patients were excluded if viral load did not promptly fall to below 50 after re-treatment during weeks 0-40.
Patients are scheduled to stop treatment at week 40 and are observed at week 52.
For 73 patients enrolled more than 52 weeks:
Those who had VL <5000 at week 52 tended to have started HAART relatively early during their infection (7% during primary HIV infection, 38% during chronic HIV infection but within 24 months after infection); had little or no rebound on study up to week 40; had low levels of proviral DNA at the start of the trial; and had lower VL before HAART.
STOP EARTH was a 52-week pilot study of STI for chronic infection in 10 patients and 20 matched controls. Treated patients were rolled-over from earlier randomized regimen comparison trials.
Baseline HIV RNA levels were above 5,000 or 10,000 copies/mL; CD4 counts were >500 cells/mm3.
Controls received no therapy and were monitored for 52 weeks. Treated patients received HAART for 52 weeks; if suppressed >20 copies/mL, patients underwent three cycles of 4 week interruptions. After the 3rd cycle, STI patients were continued off treatment and observed. Nine patients completed the study.
Patients who had spontaneous VL drops at stops 2 or 3 also had improved recovery of cytotoxic T-lymphocytes and increased CD4+ lymphoproliferative responses. With >12 months off therapy, all patients had CD4+ counts higher than pre-HAART levels.
Twenty patients received HAART with or without hydroxyurea (HU) Two-week
treatment interruptions were performed at weeks 10, 20 and 30. All patients
stopped HAART at week 40 until restart was triggered by viral load. Patients
continued hydroxyurea during the STI at week 40.
Hydroxyurea continued during STI may limit viral rebound peaks without impairing CTL and CD4 responses. STI may induce CTL and CD4 responses capable of controlling virus.
Is Treatment Interruption Safe?
This is a randomized trial in 12 patients with stable HIV RNA <400 copies/mL and CD4+ counts >400 for more than six months.
Four control patients received continuous HAART; eight patients received HAART for 4 weeks then interrupted for 4 weeks; repeated the cycle; then continued off treatment from week 16 through 28. The schedule is based on a prime/boost model.
Treatment is restarted if CD4+ counts fall to more than 50% of baseline or <200 cells/mm3; or if viral load is >5.0 log copies/mL on two consecutive visits after week 16.
Control patients maintained HIV RNA <400 copies/mL and had stable CD4+ counts. No CTL enhancement was observed.
Restoration of HIV-1-specific CTL and
Investigation of HIV Dynamics after STI in Chronic Infection
The Philadelphia Cohort
An observational study compared five untreated chronically infected patients with five chronically infected patients who interrupted treatment. Real time HIV RNA and CD4+ counts were available to their provider every two weeks to guide treatment. Restart timing was structured by scheduled visits.
During rebounds after 30-60 day interruption:
Can viral suppression be maintained with less cumulative drug exposure though a permanent on/off schedule of treatment? If intermittent therapy is effective for suppressing HIV and maintaining CD4+ cell counts, it could possibly reduce toxicity and cost and improve adherence.
A planned 22-month trial of continuous versus intermittent therapy has randomized 40 patients (70 planned) to receive either continuous HAART or to follow a cyclic regimen of HAART for one month followed by one month without treatment. All patients must have a current CD4+ T cell count > 300 cells/mm3 and a plasma HIV RNA < 50 copies/ml. The mean pre-therapy viral load for 13 patients reported through 12 weeks was 31,600 copies/mL. The mean CD4+ T cell count at enrollment was 725 cells/mm3.
Number of patients with viral load <50 copies/mL
Of 13 patients on study for 2-4 cycles, all patients had detectable plasma viremia during each of the off-HAART periods. Overall there was no difference in mean plasma viremia between the 1st and 3rd cycles off-HAART. There was a decrease in the mean CD4+ T cell count of 17, 3, and 9% with the 1st, 2nd and 3rd cycles, respectively. However, the mean CD4+ T cell count returned to baseline after 4 weeks of restarting HAART with each of the first 3 cycles.
Because all 13 patients on this schedule of intermittent therapy had detectable viral rebound while off drug, concerns were raised about possible transmissibility. A short-cycle schedule was designed to maintain rebound viral load <1500 copies during rebound. A seven-day cycle was chosen to minimize the chance of resistance and transmission due to significant replication while off therapy.
Ten patients were assigned to a six-month trial of a cyclic regimen of 7 days on HAART followed by 7 days with no treatment. Failure is considered viral load >500 copies/mL at the end of the off-drug period. To date, three patients have reached 6 months with viral load <50 copies/mL. CD4+, CD8+ and CD38+ counts were stable.
Although intermittent schedules may not reverse treatment-associated toxicity, they may lower its incidence. In this highly selected and compliant population, intermittent schedules have been well accepted by patients.
Chronic Infection: Unsuppressed Viremia
The Immunologic and Virologic Consequences
of Treatment Interruption in Clinical Practice
A clinic cohort of 75 patients on HAART who had interrupted treatment for more than 30 days was retrospectively analyzed. Patients had experience with a median of three prior regimens. The maximum pre-interruption viral load peak was a median 145,000 copies/mL; minimum pre-TI CD4+ count was a median 85 cells/mm3.
Patients had a median 573 days on HAART. This was followed by a median 67 days of interruption then by a median 171 days on HAART.
The post-TI median CD4+ count returned to baseline in 92 days; 59% recovered 90% of their pre-TI CD4+ count and 77% had viral load within 0.3 log of baseline.
STI in a Multi-drug Resistant, Unsuppressed Population
Response to re-initiation of therapy after treatment interruption was examined in the original Frankfurt STI cohort and in an expanded set of patients. In the original cohort, 33/48 patients responded with viral load < 500 copies. Twenty-four of thirty-three patients experienced rebound by a median 78 days. Those who responded with a shift to wild type virus had a more durable response.
In an expanded set of 163 patients, 104/163 (63.8%) had a virologic response. Response was associated with the change in viral load during interruption. Eighty-six percent of these had rebound during interruption. Rebound was associated with baseline CD4+ cell count, CD4 nadir, the number of drugs and the number of drug classes exposed to.
Outcome of 40 patients from the original cohort
Twenty-six of the forty patients had four or more treatment changes after ending the TI period. Thus, studies of response to re-initiation of treatment are complicated by the frequent treatment changes due to tolerability problems in this patient population.
Safety, Resistance and Efficacy of TI
in Patients with Multiple Failures of Antiretroviral Therapy Regimens
Twenty patients with a median HIV RNA of 160,000 copies/mL and CD4+ count of 77 cells/mm3 were enrolled in a pilot study to observe genotypic viral shifts during treatment interruption. The endpoint of genotypic shift was defined as the disappearance of all major resistance mutations to one drug class. Viral genotype was evaluated every 4 weeks.
After a median interruption of eight weeks, genotypic shift was observed in 11/20 patients (to one class in 5 patients; two classes in 2 patients; three classes in 4 patients).
Shifting to WT was associated with the duration of TI with shifts occurring between 8 to 10 weeks. (cf. Deeks cohort, 10-12 weeks)
Both those who experienced a shift to WT and those who did not experienced small viral load increases during TI and reductions of -2.5 log copies/mL two months after restarting treatment. By six months, the reduction was -2.0 log copies/mL. During the interruption, median CD4+ counts dropped by about 15 cells/mm3 for each group.
Clinical events (Candida esophagitis, CMV retinitis, progression of KS) occurred during interruption in patients with CD4+ counts <30 cells/mm3 at baseline.
Within three months after restarting therapy, baseline resistance mutation patterns had reemerged in 8/11 patients that shifted to WT.
Preserving CD4+ T Cells during Prolonged Virologic Failure:
A cross-sectional study of all clinic visitors during one week with CD4+ counts <350 cells/mm3 evaluated CD4 activation, Ki67 cell cycle (a proliferation assay), and cell turnover rate (k).
A subsequent study evaluated phenotypic drug susceptibility on a weekly basis in 18 heavily pretreated and drug-resistant patients who underwent treatment interruption.
A rapid return to drug susceptibility occurred in the majority of patients soon after their treatment interruption. Although the shift to wild type susceptibility was observed at different times for different patients, in each the shift was rapid when it occurred. The time of the shift was set as Day 0.
HIV RNA levels increased slowly until Day 0, then increased sharply.
CD4+ counts decreased slowly until Day 0, then decreased sharply.
Activation markers increased after interruption and this was associated with a drop in total, memory and naïve CD4 cells. CD8+ counts were stable. CD4 cell turnover increased in some patients with a decrease in the relative half-life after interruption.
Among patients with preserved CD4 cells despite virologic failure, the dynamics of T-cell production, destruction and activation are similar to those observed in treated, suppressed patients. No increase in CD4 cell production was observed after interruption.
This suggests that the wild type virus has a greater inherent fitness or replicative capacity than virus with drug resistant mutations. An increase in replicative capacity prior to week 12 was associated with increased viral load.
Patients who shift rapidly after interruption may have a less fit virus -- and therefore have the most to lose from interruption. Patients with low CD4+ counts on therapy with unsuppressed virus are at the greatest risk from disease progression during treatment interruption.
This article was provided by Project Inform. Visit Project Inform's website to find out more about their activities, publications and services.