New data on coronavirus vaccine effectiveness may be "a wakeup call"

[dalton99a]

20. Kick

The occurrence of breakthrough infections and reports of diminished neutralization of emergent variants by vaccine-elicited sera mandate the continual monitoring of the comparative effectiveness and durability of COVID-19 vaccines.8,9 Overall, we find that in our study population from Minnesota, both vaccines strongly reduce the risk of SARS-CoV-2 infection and severe COVID-19, but individuals vaccinated with mRNA-1273 were about half as likely to experience breakthrough infections as individuals vaccinated with BNT162b2. This relative risk reduction conferred by mRNA-1273 was also observed in other states, including in Florida during a recent COVID-19 outbreak. The effectiveness of both vaccines, particularly BNT162b2, was lower in July compared to prior months. Finally, the rates of complications experienced by patients with breakthrough infections were similar between those vaccinated with mRNA-1273 or BNT162b2.

mRNA-1273 and BNT162b2 were originally designed, tested, and proven to reduce the burden of symptomatic disease, hospitalization, and death related to SARS-CoV-2 infection. This study further supports the effectiveness of both vaccines in doing so, even despite the evolution of more transmissible viral variants. It is important to realize that most vaccines are not 100% effective, particularly against asymptomatic infections. For example, the estimated effectiveness of seasonal influenza vaccines has ranged from 19-60% over the past decade.16 While COVID-19 mRNA vaccines have been shown to be drastically more effective than this, the occurrence of breakthrough infections is indeed still expected. We observed a pronounced reduction in the effectiveness of BNT162b2 coinciding with the surging prevalence of the Delta variant in the United States, but this temporal association does not imply causality, and there are likely several factors contributing to changes in vaccine effectiveness over time. Consistent with our findings, a previous test-negative case-control study found that full vaccination with BNT162b2 was less effective in preventing symptomatic infection with the Delta variant (88.0%, 95% CI: 85-90.1%) than with the Alpha variant (93.7%, 95% CI: 91.6-95.3%), although it was highly effective against both.17

Several factors could contribute to the observed differences in effectiveness of mRNA-1273 and BNT162b2. Although both are nucleoside-modified mRNA vaccines encoding the prefusion stabilized SARS-CoV-2 Spike protein, there are differences in the vaccination regimen and formulation.18,19 BNT162b2 is administered as 30μg/0.3mL (100 μg/mL) doses 21 days apart20 and the Moderna vaccine is administered as 100μg/0.5mL (200 μg/mL) doses 28 days apart.21 Assuming similar sized constructs, this means that each mRNA-1273 dose provides three times more mRNA copies of the Spike protein than BNT162b2, which could result in more effective priming of the immune response. There has not been a head-to-head comparison of the neutralizing antibody titers elicited by BNT162b2 versus mRNA-1273, but such a study could provide important context for our results. Certain adverse effects, such as myalgia and arthralgia, were observed more frequently after vaccination with mRNA-1273 than BNT162b2 in their respective clinical trials, and it can be speculated that this increased reactogenicity is paralleled by increased immunogenicity.3,4 Furthermore, there are differences in the lipid composition of the nanoparticles used for packaging the mRNA content of mRNA-1273 and BNT162b2. BNT162b2 has a lipid nanoparticle composed of ALC-0315, ALC-0159, distearolyphosphatidycholine (DSPC), and cholesterol whereas the lipid nanoparticle of mRNA-1273 is composed of SM-102, PEG-DMG, DSPC, and cholesterol.22 The structures of the cationic lipids (ALC-0315 and SM-102) in each formulation are shown in Figure S5.

There are some limitations of this study. First, these cohorts are not demographically representative of the American population (Table 1, Table S1), which may limit the generalizability of our findings. Similar real world clinical studies on larger and more diverse populations from various health systems are needed to more robustly compare the effectiveness of mRNA-1273 and BNT162b2. Second, although this study accounts for geographic variability by matching individuals from the same state, these conclusions should continue to be tested longitudinally throughout the United States and globally. Third, it is possible that our vaccine effectiveness estimates are impacted by unknown exposure risk variables which were missed in the matching procedure, although the similar risks for infection, hospitalization, ICU admission, and death in the week following the first dose suggest that all of the compared cohorts had similar baseline risks for the defined outcomes at the time of study enrollment. Finally, while we did observe a recent reduction in vaccine effectiveness in July, we did not analyze the risk of infection relative to the date of vaccination. The reduced effectiveness could be due to waning immunity over time, the dynamic landscape of SARS-CoV-2 variants, or other factors that were not considered here.