Anti-neuraminidase (NA) antibodies (Ab) play a role in protection against influenza and in combination with anti-HA Ab they increase the protection in mice. To control the NA content of vaccines, which should improve vaccine standardisation and may benefit vaccine efficacy, a series of questions must be addressed: 1) The antigenic characterization of NA in vaccine strains and seed lots is based on the measurement of the enzymatic (E) activity using fetuin as substrate. The antigenic profile is established by inhibiting the E activity with post infectious ferret antisera. Overnight incubation ensures sensitivity, and fetuin substrate gives specificity by detection of variant specific antibodies. Several difficulties have to be overcome, such as the low level of E activity in MDCK grown viruses, and the lability of N1. 2) The NA protein content of the vaccines (in bulk or final product) can be measured by an ELISA capture test but the lability of the NA proteins at 4 degrees C must be checked. 3) The anti NA Ab response can be measured using a neuraminidase inhibition test. --The steric hindrance by HI antibodies does not exceed a titre of 20 in human sera. --Triton treatment of viruses reduces the steric hindrance in polyclonal sera and monoclonal antibodies but unmasks epitopes. 4) The correlations between neuraminidase inhibition, neutralization and protection, has been established in the mouse model, but remains to be shown in humans. 5) The use of a small fluorescent (MUN) or chemiluminescent (NA-STAR) substrate can be used for the rapid differentiation of N1 from N2 and NB, but not for the titration of protective NI antibodies.