Several COVID-19 vaccines use adenovirus vectors to deliver the SARS-CoV-2 spike (S) protein. Immunization with these vaccines promotes immunity against the S protein, but against also the adenovirus itself. This could interfere with the entry of the vaccine into the cell, reducing its efficacy. Herein, we evaluate the efficiency of an adenovirus-vectored vaccine (chimpanzee ChAdOx1 adenovirus, AZD1222) in boosting the specific immunity compared to that induced by a recombinant receptor-binding domain (RBD)-based vaccine without viral vector. Mice immunized with the AZD1222 human vaccine were given a booster 6 months later, with either the homologous vaccine or a recombinant vaccine based on RBD of the delta variant, which was prevalent at the start of this study. A significant increase in anti-RBD antibody levels was observed in rRBD-boosted mice (31-61%) compared to those receiving two doses of AZD1222 (0%). Significantly higher rates of PepMix™- or RBD-elicited proliferation were also observed in IFNγ-producing CD4 and CD8 cells from mice boosted with one or two doses of RBD, respectively. The lower efficiency of the ChAdOx1-S vaccine in boosting specific immunity could be the result of a pre-existing anti-vector immunity, induced by increased levels of anti-adenovirus antibodies found both in mice and humans. Taken together, these results point to the importance of avoiding the recurrent use of the same adenovirus vector in individuals with immunity and memory against them. It also illustrates the disadvantages of ChAdOx1 adenovirus-vectored vaccine with respect to recombinant protein vaccines, which can be used without restriction in vaccine-booster programs. KEY POINTS: • ChAdOx1 adenovirus vaccine (AZD1222) may not be effective in boosting anti-SARS-CoV-2 immunity • A recombinant RBD protein vaccine is effective in boosting anti-SARS-CoV-2 immunity in mice • Antibodies elicited by the rRBD-delta vaccine persisted for up to 3 months in mice.