Community Program for Clinical Research on AIDS 059 was a multicenter study conducted among human immunodeficiency virus (HIV)-infected individuals with CD4+ cell counts > or =300 cells/mm3 who were randomly assigned to receive antiretroviral therapy with or without intermittent subcutaneously administered recombinant interleukin-2 (rIL-2). To identify factors associated with a response to IL-2, a secondary analysis was performed that included the subset of rIL-2 recipients who were able to complete all 3 initial treatment cycles. Predictors of a change in CD4+ cell count between baseline and 1 month after the start of treatment cycle 3 were examined in a multivariate model that included sex, race, body surface area, rIL-2 dose, HIV load, and both baseline and nadir CD4+ cell count. The combination of race and sex (P=.027) and the nadir CD4+ cell count (P=.005) were significant predictors of mean CD4+ cell count response. Baseline CD4+ cell count had no significant effect. The strong association between nadir CD4+ cell count and CD4+ cell count response suggests that immunologic losses resulting from HIV-mediated CD4+ cell depletion may be irreversible.