Between 2013 and 2016, the A/H1N1pdm09 component of the live attenuated influenza vaccine (LAIV) produced instances of lower-than-expected vaccine effectiveness. Standard pre-clinical ferret models, using a human-like vaccine dose and focusing on antigenic match to circulating wildtype (wt) strains, were unable to predict these fluctuations. By optimising the vaccine dose and utilising clinically relevant endpoints, we aimed to develop a ferret efficacy model able to reproduce clinical observations. Ferrets were intranasally vaccinated with 4 Log10 FFU/animal (1000-fold reduction compared to clinical dose) of seven historical LAIV formulations with known (19-90%) H1N1 vaccine efficacy or effectiveness (VE). Following homologous H1N1 wt virus challenge, protection was assessed based on primary endpoints of wt virus shedding in the upper respiratory tract and the development of fever. LAIV formulations with high (82-90%) H1N1 VE provided significant protection from wt challenge, while formulations with reduced (19-32%) VE tended not to provide significant protection. The strongest correlation observed was between reduction in wt shedding and VE (R2 = 0.75). Conversely, serum immunogenicity following vaccination was not a reliable indicator of protection (R2 = 0.37). This demonstrated that, by optimisation of the vaccine dose and the use of non-serological, clinically relevant protection endpoints, the ferret model could successfully translate clinical H1N1 LAIV VE data.
Keywords: animal model; clinical translatability; ferret; influenza virus; live attenuated influenza vaccine; vaccine efficacy.