In many countries there is no clear recommendation regarding the preferred route of administration of inactivated influenza vaccines. In a randomised, observer blind study of 720 elderly subjects, a split, trivalent influenza vaccine was significantly more immunogenic for both A strains (H3N2 and H1N1, p = 0.0016 and 0.003, respectively) when given intramuscularly compared to subcutaneously. This difference was due entirely to a gender effect, with females in the intramuscular (IM) group having a significantly greater serological response than females in the subcutaneous (SC) group for both of these strains. Similar results were seen with local adverse effects. These data suggest that vaccination practices that ensure intramuscular injection are required for optimal administration of influenza vaccines in the elderly.