Background:  Temporary HAART during primary HIV-1 infection (PHI) may lower the viral setpoint, but the data from recent cohort studies are inconclusive and randomized trials have not been reported yet.

Methods:  We conducted a semi-factorial, randomized, open-label, triple arm trial in patients with PHI. A negative or indeterminate Western blot in combination with a positive p24 antigen or HIV-1 RNA test result or a negative HIV screening within the previous 180 days classified as PHI. The factorial design randomized subjects over 3 study arms (no treatment, 24 or 60 weeks of HAART), or only the 2 treatment arms if the physician or patient insisted on starting HAART. This design allowed for 2 pairwise comparisons:  between subjects who were randomized to 24 or 60 weeks of treatment, and (for subjects randomized over 3 arms) between untreated and treated patients. Early HAART consisted of zidovudine/lamivudine, efavirenz, and boosted lopinavir added until a plasma HIV RNA < 50 copies/ml was reached. Predefined endpoint was the plasma viral load at 36 weeks after seroconversion in untreated patients and 36 weeks after treatment interruption in treated patients.

Results:  At the time of analysis, 48 patients had reached the endpoint, of whom 17 were randomized over 2 arms, and 31 were randomized over 3 arms. Overall, 11 patients received no early HAART, 19 were treated for 24 weeks, and 18 were treated for 60 weeks. Mean (and 95%CI) baseline plasma viral load (10 log copies/mL) was 4.8 (4.0 to 5.6) in the untreated group, and 4.7 (4.4 to 5.0) and 5.0 (4.3 to 5.6) in the 24 and 60 weeks HAART groups. All treated patients had a plasma viral load <50 copies/mL at treatment interruption. There was no significant difference between 24 and 60 weeks of treatment in plasma viral load 36 weeks after treatment interruption (3.6 (3.0 to 4.3) vs 4.3 (3.8 to 4.8), p >0.05), but the viral setpoint was lower in early treated patients compared to untreated subjects:  4.0 (3.5 to 4.4) and 4.9 (4.7 to 5.2) log₁₀ copies/mL, respectively, p <0.005. CD4⁺ T cells at endpoint were 349 (260 to 438) cells/μL in untreated patients vs 581 (445 to 718) in treated patients, p = 0.005. Independent predictors of the plasma viral load at endpoint were baseline plasma viral load (p <0.005) and early treatment with HAART (p <0.05). At endpoint, no differences were found between the groups in proliferative capacity of Gag-specific CD4⁺ and CD8⁺ T cells, lymphocyte activation and maturation markers, or PD-1 expression.

Conclusions:  Both 24 and 60 weeks of HAART during PHI lower the plasma HIV-RNA setpoint after treatment interruption compared to no early treatment.