Comparative effectiveness of COVID-19 vaccines among health students, focusing on methodological concerns

Abstract Introduction: With emerging variants of SARS-CoV-2, there is an increasing demand for comprehensive data on vaccine effectiveness disaggregated by vaccine type and/or country to frame future pandemic readiness plans. Design, setting, participants: We investigated comparative effectiveness (VE) of CoronaVac® and Comirnaty® vaccines among health students, considering potential risk factors. An open, prospective cohort study was conducted to follow participants with valid COVID-19 vaccinations, up to 2 years. Investigations included VE against symptomatic PCR-confirmed COVID-19 infections, along with vaccine-induced humoral immunity (including durability) and potential methodologic threats to conclusive decisions. Results: Symptomatic COVID-19 incidence rate was 4.24 (95% CI = 3.69–4.86) per 10,000 person-days among 1133 students (46.6% males) over 478,466 person-days. Taking a primary series/booster with CoronaVac as the reference, a primary series with Comirnaty or a Comirnaty booster protected students up to twenty times early in the pandemic, adjusting for covariates; significance disappeared in the Omicron period, though. Unexpected upsurges in virus-specific antibody levels, starting 3-6 months after the last vaccination when titers decreased to almost nill, suggested that disproportionality in vaccine durability could have led to a bias towards the null in VE estimates, due to mediator role of undetected breakthrough infections. Conclusion: Hybrid immunity may differentially deplete the susceptibles in either arm of the study, leading to bias in VE estimations. High infection rate in Omicron period might have augmented this bias, favoring the protective effect of the less potent vaccine. Periodic PCR testing as an integrated measure in future VE studies can avoid such bias.