Twenty-eight healthy adult volunteers were immunized intranasally with an inactivated whole-virus influenza vaccine based on the strain A/New Caledonia/20/99 (H1N1), either in saline or mixed with formaldehyde-inactivated Bordetella pertussis as a mucosal adjuvant, or in a thixotropic vehicle with mucoadhesive properties. After four doses, all groups of vaccinees developed significant IgG- and IgA-antibody responses, measured by ELISA, in respectively serum and nasal secretions. None of the volunteers had demonstrable hemagglutination inhibition (HAI) antibodies in serum before being immunized, whereas more than 80% of them reached HAI titers>or=40, considered protective, after immunizations. In addition, cellular immune responses, measured as significant increases in CD4+ T-cell proliferation and granzyme B-producing cytotoxic T-cells, were detected against the vaccine strain as well as against heterologous virus strains (H3N2). However, no additive effect on these responses could be demonstrated with use of B. pertussis or the thixotropic substance in the present vaccines. It appeared, actually, that the mucoadhesive vehicle containing the thixotropic substance was less efficient than were the two other formulations. An influenza vaccine made as a simple particulate formulation of inactivated virus, and given repeatedly onto the nasal mucosa, may thus be an attractive alternative to currently available vaccines.