Background:  We sought to determine whether HAART in acute, early HIV was associated with lower HIV RNA and improved CD4 cell counts once treatment had been discontinued, compared with untreated persons.

Methods:  We used observational data from the Acute Infection Early Disease Research Program cohort. Subjects had to be enrolled within 6 months of HIV seroconversion, and chose whether to start HAART. The treatment (n = 58) group was defined as persons who started HAART within 6 months of conversion, used HAART for ≥ 12 weeks, and then stopped treatment for ≥ 4 weeks, with observations censored if treatment had resumed. The untreated (n = 337) group were persons who elected not to start HAART and had ≥ 6 months of follow-up. Primary outcomes were HIV viral load and CD4 counts at 24, 48, and 72 weeks (measured from enrollment in the untreated group and after stopping HAART in the treatment group). CD4 and viral load were modeled using a linear mixed model with natural splines that accounted for differences in baseline viral load and CD4 count. The treatment group was divided into 2 sub-groups based on when HAART was started:  ≤ 2 weeks from conversion (acute, n = 13) and > 2 weeks (early, n = 45).

Results:  Groups were similar except that the acute group had higher baseline viral load (acute median log viral load = 5.8 vs 5.1 in early treatment vs 4.9 in untreated, p < 0.01) and tended to have lower CD4 counts/mm³ (acute median CD4 = 488 vs 527 in early treatment vs 508 in untreated). The table shows differences in mean viral load and CD4 cell counts in the untreated group compared with the treatment groups by time, after adjustment for baseline differences (unadjusted results were similar though less different in significance tests).

Conclusions:  Subjects who started HAART within 2 weeks of seroconversion had viral load and CD4 count benefits that were statistically significant for 18 months after stopping HAART. In contrast, persons who started HAART between 2 and 26 weeks after seroconversion had a viral load benefit at 6 months that was not sustained at later time-points, and CD4 count benefit that appeared to decline after 12 months. These data, though non-randomized, suggest a prolonged benefit of HAART when initiated in acute HIV infection, and a more transitory benefit when treatment is initiated later in early infection.

Keywords: Primary HIV; Treatment outcomes; Antiretroviral treatment, Highly Active