This preprint reports a comprehensive meta-analysis designed to separate the contributions of immune waning and viral evolution to observed declines in COVID-19 vaccine effectiveness. The authors aimed to quantify how vaccine protection against severe disease changes with time since booster vaccination and with calendar time since the vaccine immunogen was current, using evidence from real-world studies.
The analysis integrated data identified through the ViewHub systematic review. A total of 44 studies reporting on mRNA booster vaccine effectiveness were included. The study used only openly available human data from ViewHub; the authors state that all data and code will be made available after publication.
Included boosters contained updated immunogens relevant to recently circulating SARS-CoV-2 lineages, specifically XBB.1.5, JN.1, and KP.2. The meta-analysis aggregated effectiveness estimates reported in the selected studies for these updated mRNA booster immunogens.
The authors modelled booster vaccine effectiveness while explicitly accounting for two temporal dimensions: (1) time since booster administration, interpreted as a marker of antibody waning and other within-host declines in immune protection; and (2) calendar time at which each study was conducted, interpreted as a marker of antigenic drift or viral evolution away from the vaccine immunogen. By including both time-since-vaccination and calendar-time terms, the analysis aimed to disentangle decreases in protection due to waning immunity from those due to changes in circulating viral antigenicity.
The estimated relative risk of severe disease increased by 9% per month following booster administration (95% confidence interval 7.4% to 10.5%). The authors attribute this effect primarily to immune waning.
In addition, there was an estimated increase in relative risk of severe disease of 2.4% per calendar month (95% confidence interval 1.2% to 3.6%), which the authors interpret as the effect of ongoing viral evolution (antigenic drift) relative to the vaccine immunogen.
These results indicate that booster vaccine effectiveness against severe COVID-19 declines along two distinct time axes: a relatively rapid decline after vaccination consistent with immune waning, and a slower decline over calendar time consistent with viral evolution away from the vaccine immunogen. The combined effects support the rationale for regular and timely updates of vaccine immunogens to better match circulating variants and sustain protection.
The manuscript is a preprint and has not undergone peer review. The authors used only publicly available ViewHub data; further methodological and supplementary details are referenced in the preprint and supplementary materials. All data and code will be made available after publication. Specifics about study heterogeneity, inclusion/exclusion criteria beyond the ViewHub selection, and sensitivity analyses are not detailed in the abstract; readers should consult the full preprint for those elements.
The authors declared no competing interests. They confirm adherence to ethical guidelines and note that the study used only openly available human data originally located at ViewHub. Funders named include Australian National Health and Medical Research Council Investigator Grants and the Medical Research Future Fund. The preprint was posted September 15, 2026, and bears doi: 10.64898/2026.09.14.26363077. The authors emphasize that the work has not been certified by peer review and should not be used to guide clinical practice.
From the pooled evidence of 44 studies, the authors quantify two measurable declines in booster protection against severe disease—one attributable to waning immunity after vaccination and the other to antigenic drift over calendar time. This dual-temporal decline supports policy approaches that consider both timing of booster delivery and the antigenic match of booster immunogens when designing vaccination programs.
Readers are directed to the preprint and its supplementary materials for full methods, supplementary analyses, and future availability of data and code as declared by the authors.