Fractional versus standard BNT162b2 boosters after non-mRNA priming in Mongolia: 24-month immunogenicity and safety evidence from a randomised controlled trial

Summary Background Booster doses of COVID-19 vaccines help restore protection against waning immunity and emerging variants. Having found that fractional BNT162b2 boosters were non-inferior to standard boosters in eliciting anti-spike IgG responses at day 28 in Mongolian adults, we assessed long-term immunogenicity and safety in the same cohort. Methods In this randomised, controlled, non-inferiority trial, adults previously vaccinated with two doses of ChAdOx1-S, BBIBP-CorV, or Gam-COVID-Vac were randomly assigned (1:1) to receive a 15 µg (fractional) or 30 µg (standard) BNT162b2 booster. IgG concentrations and functional surrogate virus neutralisation test (sVNT) levels against Wuhan-Hu-1 and Omicron BA.1 were assessed over 24 months. A subset had IFN-γ cell-mediated-immunity (CMI) measured (Ag1/Ag2). Immune response GMRs (fractional:standard) were estimated from log-transformed values via multivariable linear regression. SARS-CoV-2 infections, adverse events, and serious adverse events (SAEs) were recorded. ClinicalTrials.gov Identifier: NCT05265065 . Findings Of 601 participants enrolled between May 27 and Sept 30, 2022, 520 (86·5%) completed 24-month follow-up. Although IgG levels declined from six to 24 months, relative responses remained similar between arms at 18 months (GMR 1·08 [95% CI 0·97-1·22]) and 24 months (GMR 1·06 [95% CI 0·95-1·18]). In the CMI subset, IFN-γ responses peaked at day 28, waned to 18 months, and returned to baseline by 24 months, with fractional and standard arms similar at 24-month GMRs (Ag1 GMR 1·17 [95% CI 0·82-1·66] ; Ag2 GMR 1·06 [95% CI 0·73-1·54]). Median sVNT inhibition against both Wuhan-Hu-1 and Omicron BA.1 was high and comparable between groups at 18 and 24 months (both 88% [95% CI 86-90]). SARS-CoV-2 infection was confirmed in 28 participants, with an additional 386 suspected infections after day 28, inferred from a ≥1·2-fold rise in IgG titres between study visits. No SARS-CoV-2 infections resulted in hospitalisation. Fifty-three SAEs were reported, evenly distributed between groups, with no vaccine-related events. Interpretation Fractional and standard BNT162b2 boosters showed comparable neutralising activity (sVNT) that persisted to 24 months, while binding IgG declined to baseline by 24 months. In the CMI subset, IFN-γ responses followed a similar trajectory, indicating alignment of humoral and cellular immunity over time. Despite widespread SARS-CoV-2 circulation, no COVID-19 hospitalisations or deaths occurred, and safety was reassuring. These data support fractional dosing as a pragmatic, cost-saving option.