To distinguish between antigenic stimulation and CD4+ T-cell homeostasis as the cause of T-cell hyperactivation in HIV infection, we studied T-cell activation in 47 patients before and during highly active antiretroviral therapy (HAART). We show that expression of human leukocyte antigen (HLA)-DR, CD38, and Ki67 on T cells decreased during HAART but remained elevated over normal values until week 48 of therapy. We confirm previous reports that T-cell activation correlates positively with plasma HIV RNA levels (suggesting antigenic stimulation), and negatively with CD4 count (suggesting CD4+ T-cell homeostasis). However, these correlations may be spurious, because misleading, due to the well-established negative correlation between CD4 count and plasma HIV RNA levels. To resolve this conflict, we computed partial correlation coefficients. Correcting for CD4 counts, we show that plasma HIV RNA levels contributed to T-cell hyperactivation. Correcting for plasma HIV RNA levels, we show that CD4+ T-cell depletion contributed to T-cell activation. Correcting for both, activation of CD4+ and CD8+ T cells remained positively correlated. Because this suggests that CD4+ and CD8+ T-cell activation is caused by a common additional factor, we conclude that antigenic stimulation by HIV or other (opportunistic) infections is the most parsimonious explanation for T-cell activation in HIV infection. Persistence of HIV antigens may explain why T-cell activation fails to revert to levels found in healthy individuals after 48 weeks of therapy.