Immunogenicity and safety of a bivalent (omicron BA.5 plus ancestral) SARS-CoV-2 recombinant spike protein vaccine as a heterologous booster dose: interim analysis of a phase 3, non-inferiority, randomised, clinical trial

Background: SARS-CoV-2 variants evade immunity despite vaccination with prototype COVID-19 vaccines or previous infection. The 2019nCoV-311 (part 2) study is evaluating immune responses after two booster doses of a vaccine containing the omicron BA.5 subvariant spike protein in adults previously vaccinated with a prototype mRNA vaccine. This interim analysis reports on day 28 immunogenicity and safety outcomes after one booster dose. Methods: In this phase 3, randomised, observer-blinded study conducted at 35 sites in Australia, medically stable, previously COVID-19-vaccinated (mRNA-based; ≥three doses) adults aged 18 years or older were enrolled and randomly allocated (1:1:1; via an interactive web response system) to receive two doses of bivalent (NVX-CoV2373 + NVX-CoV2540; bivalent group), authorised prototype (NVX-CoV2373; prototype group), or BA.5 (NVX-CoV2540; BA.5 group) vaccine. Only blinded personnel performed study assessments or had participant contact to collect data after study vaccination. Participants received vaccines containing 5 μg SARS-CoV-2 recombinant spike protein and 50 μg Matrix-M adjuvant, administered via a 0·5 mL intramuscular injection (2·5 μg of NVX-CoV2373 plus 2·5 μg of NVX-CoV2540 for the bivalent vaccine, prepared on-site as a 1:1 mixture). The coprimary endpoints include day 28 neutralising antibody geometric mean titre (GMT) ratios (GMTRs) to omicron BA.5 and the ancestral strain, and seroresponse rates to BA.5, in the bivalent and prototype groups. These endpoints were calculated in the per-protocol analysis set, which was defined as participants who had received a vaccine dose, had baseline and day 28 immunogenicity data, and were PCR-negative for SARS-CoV-2, with no major protocol deviations. The primary objective was to determine the primary outcome (antibody responses), which consisted of three comparisons: superiority of the bivalent versus prototype vaccine for neutralising antibody GMT to BA.5 (ie, lower bound of the GMTR 95% CI >1·0); non-inferiority of neutralising antibody seroresponse rate to BA.5 (ie, lower bound of the seroresponse rate 95% CI >–5%); and non-inferiority of neutralising antibody GMT to the ancestral strain (ie, lower bound of GMTR 95% CI >0·67). This trial was registered at ClinicalTrials.gov, number NCT05372588. Findings: Between March 22, 2023 and May 2, 2023, 837 participants were screened for eligibility and 766 were randomly allocated to receive the BA.5 (n=255), prototype (n=252), or bivalent (n=259) vaccine. After accounting for exclusions due to participants being baseline SARS-CoV-2-positive, having previous infection, or protocol deviations, the per-protocol analysis set included 694 participants (236 in BA.5 group, 227 in prototype group, and 231 in bivalent group). In this interim analysis (maximum follow-up 35 days after the first dose), the bivalent group, compared with the prototype group, had superior neutralising antibody responses to BA.5 (GMT 1017·8 [95% CI 891·0–1162·6] vs 515·1 [450·4–589·0]; GMTR 2·0 [1·69–2·33]) and a non-inferior seroresponse rate to BA.5 at day 28 (39·8% [33·5–46·5] vs 12·3% [8·4–17·3]; difference 27·5% [19·8–35·0]). The bivalent group also had non-inferior neutralising antibody responses to the ancestral strain (GMTR 1·0 [0·84–1·20]), compared with the prototype group. All vaccines were similarly well tolerated. Interpretation: All three coprimary endpoints were met in part 2 of the ongoing 2019nCoV-311 study. These data support the development of monovalent and/or bivalent vaccines for the most currently circulating variants, to optimise protection. With no new safety findings, further investigation of omicron-based subvariant vaccines is supported by the evidence. Funding: Novavax.