While the original recombinant HIV vaccines used vaccinia virus and were well tolerated, safety considerations in immunodeficient hosts has led investigators to substitute the canarypox virus for the vaccinia virus. However, the immunogenicity of the pox-virus-based vaccines in humans has been relatively modest with less than 35% of the vaccinees scoring positive for T-cell responses. In contrast, replication-defective adenovirus type 5 (ad5) appears to be one of the most promising live virus vectors for HIV vaccines. Merck has used this complex and showed that 50% of the volunteers had significantly longlasting CD8 T-cell responses to HIV-1 peptides. A trivalent recombinant ad5-gag/pol/nef complex has been engineered and retested in human volunteers. A phase II trial of this candidate is being investigated in 1,200 men and 400 women at high risk of exposure to HIV infection who will be followed for three years.