Three cohort studies in adults were performed during the period from 1986 to 1989. Eight hundred and eighty-four subjects were, one or more times, immunized with influenza vaccines, and pre- and post-vaccination antibody titres were determined by hemagglutination inhibition tests. One thousand and one hundred and nineteen vaccination events in 681 subjects could be analysed by a comparison, per trial and per influenza (sub)type, between groups with and without influenza vaccination in previous years. Effect size, odds ratio and protection rate difference, were used as effect measures. Subjects with previous vaccination showed higher pre-vaccination antibody than those without. The average change of the post-vaccination proportion of subjects with high antibody titre value to previous vaccination, was +9.4% (95% CI: +5.3 to 13.6%) for A-H3N2 vaccine components, -2.1% (-8.1 to 3.9%, not significant) for A-H1N1 and -10.6% (-16.5% to -4.8%) for B. In a linear regression model, pre-vaccination titres and the status of previous vaccination were identified as factors significantly influencing post-vaccination titres. These findings are discussed in the context of a short review of the literature. It is concluded that the status of previous vaccination should always be addressed as an independent factor in serological vaccination studies.