Rationale for two influenza B lineages in seasonal vaccines: A meta-regression study on immunogenicity and controlled field trials.

B lineage mismatch prompted introduction of quadri-valent influenza vaccines (QIV) with two influenza B viruses representing distinct antigenic lineages. To explore the impact on antibody induction and vaccine effectiveness predicted from antibody (VEab), we performed a systematic literature search on immunogenicity studies conducted to assess antibody superiority of QIV over trivalent influenza vaccine (TIV). Thirteen relevant articles described 31 trials from 2007 and 2013. Log-transformed GMT trial estimates and their variances were converted to clinical protection rates predicted from antibody (PRab). VEab estimates were calculated from pre- and post-vaccination PRab. Without specific pre-vaccination immunity, average VEab was 69% for match, and -4% for lineage mismatch. With increasing pre-vaccination seropositivity, mismatch impact declined to 2%. We also performed an umbrella literature search for randomised controlled trials and test-negative case-control trials with TIV, and estimated vaccine effectiveness against laboratory-confirmed influenza B (VEf). Sixty-eight eligible clinical articles described 110 season-trials from 1965 to 2012, covering seasons with B lineage match (n=52), lineage drift (n=15) and lineage mismatch (n=43). With no pre-vaccination antibody levels determined, we used chance of previous exposure to influenza B (Ppe) as pre-seasonal immunity measure. When Ppe was 0%, average VEf for matched seasons was 67%, and for mismatched seasons 35%, indicating a moderate, yet significant mismatch impact on VEf. With increasing Ppe, mismatch impact declined to 3%. Thus serological and field trials indicate that B lineage mismatch impact is negatively related to pre-seasonal immunity and that the gain of QIV over TIV most benefits infants and children not yet exposed to influenza B.