“Authorized COVID-19 vaccines train the immune system with noninfectious components and protect especially against severe COVID-19.”

• None of the vaccines authorized in the U.S. (Pfizer-BioNTech/Comirnaty, Moderna/Spikevax, Novavax/Nuvaxovid, and the discontinued Janssen) contain live, replication-competent SARS-CoV-2. mRNA vaccines deliver lipid-nanoparticle-encapsulated mRNA coding for the spike protein; Novavax delivers a recombinant spike protein nanoparticle with a saponin adjuvant; the Janssen vaccine used a replication-incompetent adenovirus 26 vector. All deliver antigen (or antigen instructions) that teach B and T cells to recognize spike without the ability to cause COVID-19 — exactly the 'noninfectious components' description (CDC, FDA, WHO).
• Pivotal randomized trials showed high efficacy against disease including severe disease: BNT162b2 ~95% efficacy against symptomatic COVID-19 (Polack et al., NEJM 2020); mRNA-1273 94.1% efficacy with all observed severe cases in the placebo group (Baden et al., NEJM 2021); NVX-CoV2373 ~90% efficacy with 100% against moderate-to-severe disease in the PREVENT-19 trial (Dunkle et al., NEJM 2022).
• Post-authorization surveillance consistently finds that protection against severe outcomes (hospitalization, ICU admission, death) is substantially higher and more durable than protection against infection, including during Delta and Omicron waves — the central asymmetry the claim asserts. CDC MMWR analyses of emergency-department and hospitalization data and UKHSA vaccine surveillance reports both document this pattern.
• A systematic review and meta-regression in The Lancet (Feikin et al., 2022) found vaccine effectiveness against severe disease declined only modestly (about 10 percentage points over 6 months) compared with larger declines against infection — i.e., the protective effect is preferentially retained against severe COVID-19.
• CDC rate-by-vaccination-status surveillance repeatedly showed markedly higher COVID-19 death and hospitalization rates among unvaccinated versus vaccinated/boosted people across multiple variant periods.

• 'Noninfectious components' is a simplification for viral-vector vaccines: the Janssen/J&J and AstraZeneca products contain actual adenovirus particles. They are replication-incompetent and cannot cause infection with adenovirus or SARS-CoV-2, but they are whole (if crippled) virus particles rather than purely inert molecules. The same nuance applies to whole-inactivated vaccines authorized outside the U.S. (Sinovac, Sinopharm, Covaxin) — killed virus, noninfectious, but derived from the whole pathogen.
• 'Protect especially against severe COVID-19' is a relative, population-level statement. Protection is not absolute: breakthrough severe disease and death occur, particularly in elderly and immunocompromised people, and effectiveness against severe outcomes does wane measurably over months and with antigenically drifted variants, requiring updated boosters.
• Efficacy estimates against severe disease in the original trials rested on small numbers of severe endpoints, so the confidence intervals were wide; the stronger evidence is observational and therefore subject to confounding (e.g., healthy-vaccinee effects), though the direction of the effect is corroborated across many independent datasets.
• The claim is silent on rare harms (myocarditis/pericarditis after mRNA vaccines in young males; thrombosis with thrombocytopenia syndrome after adenovirus-vector vaccines, which led FDA to limit and ultimately withdraw the Janssen vaccine). This does not contradict the claim as written but is relevant omitted context.

93/100 — the lab’s own confidence in the score above, not a second reading of the claim.

U.S. Centers for Disease Control and Preventiongov · 92 · New England Journal of Medicine (Polack et al., BNT162b2 trial)journal · 96 · New England Journal of Medicine (Baden et al., mRNA-1273 trial)journal · 96 · New England Journal of Medicine (Dunkle et al., NVX-CoV2373 PREVENT-19 trial)journal · 95 · The Lancet (Feikin et al., duration of vaccine effectiveness meta-regression)journal · 94 · CDC MMWR (Ferdinands et al., waning 2- and 3-dose mRNA effectiveness against ED visits and hospitalization)gov · 92 · UK Health Security Agencygov · 90 · World Health Organizationgov · 90 · Johns Hopkins Medicineedu · 88

• Government and public-health sources consistently describe authorized COVID-19 vaccine platforms as not containing infectious SARS-CoV-2: mRNA vaccines provide genetic instructions for a spike-protein immune response, protein-subunit vaccines contain harmless pieces of the virus, and inactivated/viral-vector approaches are designed so they cannot cause COVID-19.
• The claim that vaccines 'train the immune system' is well supported: CDC and WHO describe vaccination as exposing the immune system to an antigen or instructions for making one so immune memory can respond faster to later SARS-CoV-2 exposure.
• Clinical and real-world evidence supports the 'especially against severe COVID-19' part. CDC MMWR data found mRNA vaccination was highly effective against COVID-19-associated invasive mechanical ventilation and death, though effectiveness was lower during Omicron than before.
• A Lancet systematic review/meta-regression found protection against severe COVID-19 was higher and more durable than protection against infection, matching the claim’s emphasis on severe disease.

• The statement should not be read as meaning vaccines completely prevent infection or transmission. WHO and CDC note that vaccinated people can still get COVID-19, and no vaccine is 100% protective.
• Protection changes over time and by variant. The Lancet review found vaccine effectiveness wanes, especially against infection, and a NEJM Qatar study found limited two-dose protection against symptomatic Omicron infection months after vaccination, although protection against severe disease remained stronger.
• The phrase 'noninfectious components' is a simplified description. Some authorized vaccines use nonreplicating viral vectors or genetic instructions rather than simply a physical 'component' of SARS-CoV-2; however, they are still not infectious SARS-CoV-2 and cannot cause COVID-19.

96/100 — the lab’s own confidence in the score above, not a second reading of the claim.

Centers for Disease Control and Preventiongov · 95 · World Health Organizationgov · 94 · CDC Morbidity and Mortality Weekly Reportgov · 92 · The Lancetjournal · 95 · New England Journal of Medicinejournal · 95

• Authorized COVID-19 vaccines (including mRNA, viral vector, and protein subunit types) do not contain the live SARS-CoV-2 virus. They utilize noninfectious genetic instructions (mRNA) or harmless viral proteins to teach the immune system to recognize and attack the virus.
• Decades of vaccine research and massive real-world data tracking since the rollout of COVID-19 vaccines confirm that their primary and most durable benefit is preventing severe outcomes. While protection against mild symptomatic infection can wane as the virus mutates, cellular immunity (T-cells) remains highly effective at preventing severe disease, hospitalization, and death.

• There is no credible scientific or medical evidence against the claim as written. The assertion perfectly aligns with global immunological consensus.
• The only counter-arguments stem from a misinterpretation of vaccine efficacy, where detractors point to 'breakthrough infections' as proof vaccines do not work. However, the claim specifically and accurately states that the vaccines protect 'especially against severe COVID-19,' which correctly accommodates the reality of breakthrough mild infections.

100/100 — the lab’s own confidence in the score above, not a second reading of the claim.

Johns Hopkins Medicineedu · 98 · Yale Medicineedu · 95

• WHO states that authorized COVID-19 vaccines use noninfectious material: protein-based vaccines contain harmless protein fragments or shells, while mRNA vaccines provide genetic instructions that prompt cells to produce an immune-stimulating protein; neither approach requires live SARS-CoV-2 infection.[2]
• WHO states that all COVID-19 vaccines it lists for emergency use or prequalification protect against severe disease and death, though they are less effective against mild disease.[2]
• A Lancet systematic review found that vaccine effectiveness against severe disease generally remained above 70% after primary vaccination, with only modest decline through about six months.[1]
• A large Lancet cohort study found 90% effectiveness against COVID-19-related hospital admission after full BNT162b2 vaccination, although effectiveness against infection declined substantially over time.[3]
• A Swedish national cohort study found approximately 89% effectiveness against severe COVID-19 shortly after vaccination and 64% effectiveness after 121 or more days, indicating stronger and more durable protection against severe outcomes than against infection.[7]

• The wording "noninfectious components" is broadly accurate but oversimplifies vaccine platforms: some vaccines deliver mRNA or other biological instructions rather than merely presenting inert components, although they do not contain live virus capable of causing COVID-19.[2][8]
• Protection is not absolute. WHO notes that vaccinated people can still become infected and develop COVID-19, particularly because protection against mild disease is lower.[2]
• Protection wanes. The Lancet review found a decline of about 10 percentage points in effectiveness against severe disease, and the Swedish study found effectiveness against severe disease falling from 89% to 64% after roughly four months.[1][7]
• The claim does not specify vaccine, variant, dose schedule, time since vaccination, age, or immune status; these factors affect protection, so "protect especially" is directionally correct but not a universal numerical guarantee.[1][3][7]

96/100 — the lab’s own confidence in the score above, not a second reading of the claim.

World Health Organizationgov · 98 · The Lancetjournal · 97 · The Lancetjournal · 97 · The Lancetjournal · 97

• Authorized COVID-19 vaccines (mRNA, non-replicating viral vector, and protein subunit) contain no live/infectious SARS-CoV-2; they present spike protein or genetic instructions for it so the immune system can generate antibodies and T-cell responses.
• Large randomized trials (Pfizer-BioNTech, Moderna, others) showed high efficacy against symptomatic disease and still higher efficacy against severe COVID-19, hospitalization, and death.
• Subsequent observational studies, CDC analyses, and WHO summaries consistently find that remaining protection is strongest against severe outcomes even after infection-blocking immunity wanes with time and variants.

• Effectiveness against infection and mild disease is lower and declines faster than effectiveness against severe disease (the claim already qualifies this with 'especially').
• Inactivated whole-virus products used in some countries are noninfectious after processing but are not 'subunit components'; they were not the primary authorized products in the US/EU. These nuances do not falsify the statement as written.

90/100 — the lab’s own confidence in the score above, not a second reading of the claim.

New England Journal of Medicinejournal · 98 · World Health Organizationother · 92 · CDCgov · 95