Bordetella pertussis causes whooping cough (pertussis), a severe and sometimes fatal respiratory infectious disease, especially in young infants. Pertussis can be prevented in infants and children by immunization with either whole-cell pertussis (wP) or acellular pertussis (aP) vaccines; however, its incidence is increasing in many countries despite high vaccine coverage. This resurgence in populations immunized with aP vaccines has been attributed to (1) genetic changes in circulating strains of B. pertussis resulting from vaccine-driven immune selection, (2) waning protective immunity due to poor induction of immunological memory, or (3) a failure of aP vaccines to induce the appropriate arm(s) of the cellular immune responses required to prevent infection. Studies in a baboon model have suggested that previous infection prevents reinfection as well as disease, whereas aP vaccines fail to prevent nasal colonization and transmission of B. pertussis. Studies in the mouse model have demonstrated that immunization with wP vaccines induces Th1 and Th17 responses, whereas aP vaccines promote Th2-skewed responses and high antibody titres. Thus, while aP vaccine-induced antibodies may prevent pertussis, they may not prevent nasal colonization or transmission. Emerging data have suggested that replacing alum with novel adjuvants based on pathogen-associated molecular patterns has the capacity to switch the responses induced with aP vaccines to the more protective Th1/Th17 responses and may also enhance immunological memory. It is likely that third-generation pertussis vaccines will be based on live attenuated bacteria or aP formulations with novel adjuvants, which prevent nasal and lung infection and induce sustained immunity through induction of memory T cells.