Adult Clinical Trials Group Study 349 examined the immunology, virology, and safety of 40 mg/d prednisone as an adjunct to antiretroviral therapy in 24 HIV-infected subjects with >200 CD4+ T cells/mm in a randomized placebo-controlled trial. After 8 weeks, median lymphocyte and CD4+ cell numbers increased >40% above baseline values (p =.08). No effect was observed on markers of cell activation or apoptosis, although the proportion of CD28+ CD8+ T cells increased (p =.006). Prednisone inhibited monocyte TNFalpha production without affecting T-cell responses to antigens or mitogens. Two subjects assigned to prednisone were subsequently found to have asymptomatic osteonecrosis of the hip. Many questions remain regarding the role of activation-induced sequestration and apoptosis as causes of progressive CD4+ T-cell loss in AIDS. The potential role of corticosteroids as tools to examine this question will be limited by concerns regarding their toxicity; however, further studies of other agents to limit cellular activation in AIDS are warranted.