Randomized, controlled trial of therapy interruption in chronic HIV-1 infection

Abstract

Background: Approaches to limiting exposure to antiretroviral therapy (ART) drugs are an active area of HIV therapy research. Here we present longitudinal follow-up of a randomized, open-label, single-center study of the immune, viral, and safety outcomes of structured therapy interruptions (TIs) in patients with chronically suppressed HIV-1 infection as compared to equal follow-up of patients on continuous therapy and including a final therapy interruption in both arms.

Methods and findings: Forty-two chronically HIV-infected patients on suppressive ART with CD4 counts higher than 400 were randomized 1:1 to either (1) three successive fixed TIs of 2, 4, and 6 wk, with intervening resumption of therapy with resuppression for 4 wk before subsequent interruption, or (2) 40 wk of continuous therapy, with a final open-ended TI in both treatment groups. Main outcome was analysis of the time to viral rebound (>5,000 copies/ml) during the open-ended TI. Secondary outcomes included study-defined safety criteria, viral resistance, therapy failure, and retention of immune reconstitution. There was no difference between the groups in time to viral rebound during the open-ended TI (continuous therapy/single TI, median [interquartile range] = 4 [1-8] wk, n = 21; repeated TI, median [interquartile range] = 5 [4-8] wk, n = 21; p = 0.36). No differences in study-related adverse events, viral set point at 12 or 20 wk of open-ended interruption, viral resistance or therapy failure, retention of CD4 T cell numbers on ART, or retention of lymphoproliferative recall antigen responses were noted between groups. Importantly, resistance detected shortly after initial viremia following the open-ended TI did not result in a lack of resuppression to less than 50 copies/ml after reinitiation of the same drug regimen.

Conclusion: Cycles of 2- to 6-wk time-fixed TIs in patients with suppressed HIV infection failed to confer a clinically significant benefit with regard to viral suppression off ART. Also, secondary analysis showed no difference between the two strategies in terms of safety, retention of immune reconstitution, and clinical therapy failure. Based on these findings, we suggest that further clinical research on the long-term consequences of TI strategies to decrease drug exposure is warranted.

Conflict of interest statement

Competing Interests:

RMG is a paid consultant for the Bayer Guidelines Project, which develops algorithms for interpretation of drug resistance genotyping assays; received honoraria and research support from ViroLogic and Visible Genetics; received honoraria for speaking at educational programs supported by ViroLogic, Visible Genetics, GlaxoSmithKline, Bristol-Myers Squibb, Roche Pharmaceuticals, and Agouron Pharmaceuticals; and directs a nonprofit academic laboratory at the Gladstone Institute of Virology and Immunology that has provided services for clinical research supported by grants to the University of California from Merck, GlaxoSmithKline, Bristol-Myers Squibb, Boehringer Ingelheim, Roche, Abbott, Agouron Pharmaceuticals, Gilead, Visible Genetics, and Chiron.

RG receives support for his HIV research from Agouron Pharmaceuticals, GlaxoSmithKline, and Bristol-Myers Squibb and serves as a consultant to GlaxoSmithKline; all relationships have been disclosed to the University of Pennsylvania, which deemed them not to constitute a conflict of interest.

Figures

Figure 1. Study Flow

Figure 2. Study Design (Phases I and II)

Figure 3. Lack of a Difference between Groups in Plasma HIV-1 RNA during Phase II

Top panel shows Kaplan-Meyer plot summarizing time to a viral load of more than 5,000 copies/ml in both arms. Second panel shows viral load (mean ± standard error) per arm during 27 wk of TI (median time of phase II). Bottom table shows number of patients at time points shown for viral load in the second panel; the decrease in viral load over time is due to the reinitiation of therapy in patients with higher viral loads.

Figure 4. T Cell Subsets and Recall Lymphoproliferative Response at the End of Phase I

End of phase I values for each arm are summarized (median and first and third quartiles) in the stacked figures showing from top to bottom: CD4 T cells/μl, CD4%, CD4−CD45RA+CD62L+% (naïve phenotype), CD8 T cells/μl, CD8%, and C. albicans lymphoproliferative response (shown as stimulation Index, SI). Unpaired p values for each variable are shown above corresponding bracket.

Figure 5. CD4 T Cells/μl and T Cell Recall Lymphoproliferative Response during Sequential TIs in Phase I

Shown are data from the repeated interruptions arm. Panels show the TI initiation visit and TI end visit of each sequential TI inclusive of the initiation visit for phase II (open-ended TI).