Thus, antibodies to the conserved region of influenza HA could potentially be involved in the observed heterosubtypic protection. To assess the immune responses of the trial CPI-613 participants in vitro, we used the HAI, VNA and ADCC assays. These commonly used assays were selected because each measures a different mechanism of viral neutralization and their combination presumably enables quantification of broad and effective humoral immune response. Serum titers of antibodies measured with all three in vitro assays correlated with protection, as characterized by survival and reductions in bodyweight loss and clinical scores, following homologous H1N1 challenge. Thus, we confirm that HAI, which is a known in vitro correlate of protection against homologous influenza in humans, is also identified as a correlate of homologous protection in passive transfer. Only ADCC titers correlated significantly with protection against H5N1. Whereas the serum-mediated H5N1 protection and its transient nature were revealed using the serum transfer and challenge model, these effects would have been missed based on results of the in vitro assays only. These findings underscore that different antibodies with different mechanisms of action play a role in protection against different influenza strains. As a consequence, no single in vitro assay may accurately assess the protective activity, particularly of novel vaccine modalities designed to confer broad protection. Our novel serum transfer and challenge model can be used to further dissect the humoral components that contribute to protection. Mouse passive transfer models have been used previously to assess the protective activity of human serum samples. However, the design applied here in which each individual human serum sample is transferred to an individual mouse allows for the evaluation of a substantial number of serum samples in a single challenge study while retaining sufficient statistical power. Furthermore, this design enables not only the evaluation of protective responses per group and per time point but also monitoring of changes in the immune response and protective ability of serum from individual donors at different time points. In this study, the humane endpoint following challenge was – similar to the mouse LD50 determinations for both challenge viruses – defined on a clinical score as a measure of discomfort, rather than on a certain percentage of bodyweight loss. However, other humane endpoints will be equally suited for the read out in serum transfer challenge experiments.