Background:  Induction therapy followed by treatment interruption has been examined as a strategy in patients identified in primary HIV infection. We analyzed whether pre-treatment clinical, virological, or immunological variables were associated with more rapid viral suppression on therapy, and subsequently delayed viral rebound off therapy.

Methods:  In a randomized trial (PULSE), both acute (defined as £ 3 Western blot bands and p24 antigenemia) and early primary HIV infection (seroconversion within a 6-month period) patients were treated with combination therapy, with or without hydroxyurea. After 6 to 12 months, if viral load was <50 copies/mL, therapy was interrupted. Factors associated with median time to initial viral suppression and median time to viral rebound (>5000 copies/mL) during treatment interruption, were determined using a Cox proportional hazards model, stratified (blinded) for randomisation. Baseline variables included: age, acute/early status, acute retroviral syndrome, viral load, CD4 count. Markers of T-cell activation were also included for a subset of patients.

Results:  In the first treatment phase, 68 male patients were assessed, 28 acute and 40 early primary HIV infection. Mean age 35 years and median viral load 5.73 log₁₀ (IQR:  4.82 to 5.87). The median time to <50 copies/mL was 113 days (IQR:  84 to 168 days). In multivariate analyses, a low baseline (pre-treatment) viral load (<50,000 copies/mL) was an independent predictor (p = 0.009) of a shorter time to suppression. In a subset of patients (n = 33), after adjustment for baseline RNA, increased baseline CD28⁺% of CD8⁺ T cells (>45%) was also significantly (p = 0.013) associated with a shorter time to suppression. Following interruption in 58 subjects, the median time to rebound (>5000 copies/mL) was 30 days, (IQR:  22 to 106 days). Predictive of a longer time to rebound was a lower baseline viral load (<50,000; p = 0.041), and a shorter time to viral suppression on therapy (<16 weeks; p = 0.027). In a subset of patients (n = 34), baseline CD28⁺% of CD8⁺ T cells (>45%) was only significantly associated with longer time to viral rebound (p = 0.016) in univariate analyses.

Conclusions:  These results show that, during primary HIV infection a lower viral load at initiation of therapy was associated with more rapid viral suppression and a delayed rebound. Furthermore, maintenance of a central pool of CD28⁺% CD8⁺ T cells (with proliferative potential), rather than expansion of CD28-effector or CD38⁺ activated CD8⁺ T cells, was also associated with better control on therapy.

Keywords: Primary HIV Infection; Treatment; Interruption