Here we investigate the effect of viremia and the influence of HAART on the frequency and quality of HIVspecfic T cells in an adolescent/young adult cohort. Measurements of viral loads and the magnitude and quality of antiviral cellular immune responses were performed on 14 HAART-naive and 8 treated HIV-1-infected adolescents. Cross-sectional correlations between viral load and cellular immune responses were determined and data were analyzed by viral load (<4000, 4000-40,000, and >40,000 copies/ml plasma) and patient treatment status. All 22 patients showed a broad IFN-gamma ELISPOT response that was proportional to viral load (r = 0.53, p = 0.02), recognizing an average of five to eight peptide pools throughout Gag, Pol, Env, Tat, Rev, and Nef. Intracellular cytokine staining was performed with pools of overlapping peptides corresponding to HIV Gag to distinguish CD8 response from CD4 response. Among untreated patients with increased viral load there was a constant IFN-gamma CD8 response but a declining IFN-gamma CD4 response. HIV-specific IL-2 production was consistently low in CD8 cells but inversely related to viral load in CD4 cells (r = -0.52, p = 0.02). In this crosssectional analysis, time on HAART was associated with an increased frequency of antiviral IFN-gamma- and IL-2-coproducing CD4 cells (r = 0.98, p <0.001), but not of antiviral CD8 cells. Our results suggest that T cells coproducing IL-2 and IFN-gamma are a better marker for immunological competence than T cells producing IFN-gamma alone. They also suggest that HAART may be associated with an improved capacity for IL-2 production by antiviral CD4 T cells in a time-dependent manner. Longitudinal studies are clearly necessary to assess the impact of HAART on these parameters.