Background:
The evaluation of strategies targeting HIV cure requires analytical antiretroviral treatment (ART) interruption (ATI) to measure the effect on viral load dynamics with the ultimate goal of maintaining viral control in the absence of ART. The objective of this study was to evaluate potential viral endpoints and the impact of the trial design using 3 ATI trials.
Methods:
The ANRS 118 ILIADE is a phase 2 trial of 148 patients randomized to either IL-2 subcutaneous therapy (n=81) or placebo (n=67) and followed during 228-week ATI period. The VRI02 ANRS 149 LIGHT is a phase 2 trial of 103 patients randomized to either recombinant DNA vaccine (GTU-MultiHIV B) followed by lipopeptide vaccine (HIV-LIPO-5) or placebo and followed during a 12-week ATI period. The ANRS/VRI DALIA is a single arm phase I trial evaluating the effects of a dendritic cell vaccine in 19 infected patients followed over a 24-week ATI period. Various HIV RNA endpoints were calculated in each trial: time to rebound (TTR), the maximum observed value (peak), the slope of the initial increase of HIV RNA (slope), the area under the curve (AUC) and the stable level reached after rebound (setpoint).
Results:
The median TTR, peak, slope, AUC and setpoint varied across arms from 2 to 6 weeks, 4.6 to 5 log10 copies/mL, 0.22 to 0.75 log10 copies/mL/week, 3.74 to 4.25 log10 copies/mL and 4.29 to 4.59 copies/mL, respectively. These endpoints were sensible to the design of the trial (sampling times and duration of ATI).When estimating the correlations between endpoints within each trial, it appears that there was no correlation between the setpoint and the TTR or the slope (but in LIGHT vaccine arm) whereas the maximum correlation was always estimated between setpoint and AUC (r between 0.79 and 0.99, all p< .0001).Then, we simulated the impact of the threshold of HIV RNA defined to resume ART varying from 2 to 6 log10 copies/mL and whether a confirmatory measure is asked when the threshold is passed. The correlations between the AUC and the setpoint were the highest for a threshold of 5 log10 copies/mL and with a confirmatory measure (r >0.8).
Conclusions:
The evaluation of time to rebound requires frequent monitoring of HIV RNA (at least once a week over 6 weeks) and does not reflect at all the viral setpoint in case of viral replication. For the evaluation of the capacity to control viral replication, we recommend to use of the AUC where ART are resumed if HIV RNA goes above 5 log10 copies/mL.