The induction of cell-mediated immunity (CMI) to Bordetella pertussis antigens (whole, heat-inactivated bacterial cells [BPC], pertussis toxin [PT], filamentous haemagglutinin [FHA], pertactin [PRN]) was assessed by a lymphoproliferation assay in vitro in a cohort of children enrolled in a randomized clinical trial of pertussis vaccines efficacy in Italy. Four vaccination groups were compared: children receiving acellular pertussis (aP) vaccines from SmithKline Beecham (SB) or Chiron Biocine (CB) or whole-cell vaccine (wP) from Connaught, each combined with diphtheria and tetanus toxoids (DT), or a DT vaccine only. When the purified antigens were used, statistically significant differences in CMI responses were observed between pre- and post-vaccination samples. In particular, CMI responses to FHA and PRN were detected in the majority of both aP vaccines recipients, whereas DTwP-recipients were CMI-positive in a much lower proportion. Clear-cut differences in PT responses were detected between DTwP and DTaP vaccine recipients, in favour of the latter. These differences were maintained up to 24 months after completion of the primary vaccination schedule. Thus, CMI responses could be a useful adjunct to serology in studying the immune responses to pertussis vaccines.