397    

Background: In the context of HAART regimen interruption, our objectives were:   to estimate viral load within the year following interruption of HAART initiated among 58 primary HIV-1-infected subjects enrolled in the prospective PRIMO Cohort; to compare this viral load with viral load reached spontaneously at the same delay since infection in natural history among 116 seroconverters followed in the prospective SEROCO Cohort between 1988 and 1996.

Methods:  Since 1996, the French ANRS PRIMO Cohort has enrolled 393 HIV⁺ patients diagnosed during primary infection confirmed by: incomplete Western blot, or a positive p24 antigenemia, or an ELISA test + within 6 months. Our study patients had to:  initiate HAART within 6 months after infection;  be virological responders on HAART (i.e. <500 copies/mL between the first response and the treatment interruption). A linear mixed-effects model with 2 slopes was used to describe the dynamics of viral load after interruption among the 58 patients, using a total of 220 viral load measurements carried out during interruption. The time threshold was estimated by profile likelihood. Among the 116 patients enrolled within 6 months following infection in the SEROCO Cohort, viral load level in natural history was estimated 36 months since infection, within a window of 3 months.

Results:  Viral load reached 3.57 log copies/mL at 27 days after interruption (estimated time threshold); the second slope was 0.03 log copies/mL per month, which leads to an estimation at 12 months of 3.94 log copies/mL. This viral load was associated with both viral load at HAART initiation and CD4⁺ cell count at interruption. HAART precocity (0 or 1 band on Western blot; delay since first symptoms <3 weeks; delay from infection <1.5 month) and duration of the sustained response to HAART before interruption were not associated with viral load during interruption. The estimated viral load reached after a 24-month treatment and a 12-month interruption was 3.75 log copies/mL (adjusted for the proportion of females and viral load at entry observed in the SEROCO Cohort) versus viral load spontaneously reached 36 months since infection in the SEROCO cohort: 3.94 log copies/mL.

Conclusions:  We did not find evidence of strong benefits of early treatment on viral load after interruption of effective HAART regimen. Randomized trials with long-term follow-up are needed to assess long-term potential benefits of interruption of early and effective HAART.

Keywords: HAART interruption; Viral load; primary infection