Abstract Background Dose fractionation of a coronavirus disease 2019 (COVID-19) vaccine could effectively accelerate global vaccine coverage, while supporting evidence of efficacy, immunogenicity, and safety are unavailable, especially with emerging variants. Methods We systematically reviewed clinical trials that reported dose-finding results and estimated the dose-response relationship of neutralizing antibodies (nAbs) of COVID-19 vaccines using a generalized additive model. We predicted the vaccine efficacy against both ancestral and variants, using previously reported correlates of protection and cross-reactivity. We also reviewed and compared seroconversion to nAbs, T cell responses, and safety profiles between fractional and standard dose groups. Results We found that dose fractionation of mRNA and protein subunit vaccines could induce SARS-CoV-2-specific nAbs and T cells that confer a reasonable level of protection (i.e., vaccine efficacy > 50%) against ancestral strains and variants up to Omicron. Safety profiles of fractional doses were non-inferior to the standard dose. Conclusions Dose fractionation of mRNA and protein subunit vaccines may be safe and effective, which would also vary depending on the characteristics of emerging variants and updated vaccine formulations.