Is covid vaccine safe

The short version: the evidence on COVID-19 vaccine safety is deeply conflicted, and the studies that show net benefit are all observational, not randomized, so they cannot separate the effect of the vaccine from the effect of being a healthier person who chooses to vaccinate.

The strongest evidence for benefit comes from large observational studies. A Norwegian cohort of 4.6 million adults found that people with three or more mRNA doses had about 58% lower all-cause mortality than the unvaccinated [6]. A Hungarian study of 6.4 million adults, which tried to correct for the "healthy vaccinee" bias by using a non-epidemic period as a control, still found a 49-75% reduction in all-cause death depending on the vaccine [7]. A Danish study of 102,000 heart failure patients found a small but statistically significant 0.3% absolute reduction in 90-day mortality after mRNA vaccination [13]. These are real signals, but every one of them is observational. The vaccinated and unvaccinated differ in ways that cannot be fully adjusted for: the unvaccinated group includes people too sick to vaccinate, people with vaccine contraindications, and people who distrust the system for reasons that correlate with other health risks.

The evidence for harm is also substantial and comes from multiple independent analyses. A UK analysis of ONS data from April 2021 to May 2023 found that all-cause mortality in vaccinated individuals started lower than the unvaccinated but crossed over and became significantly higher by late 2021, with the gap widening over time [5]. A US county-level study of over 3,000 counties found that higher vaccination uptake was associated with higher all-cause excess mortality in both 2022 and 2023 [3]. A European ecological study of 31 countries found that a one-percentage-point increase in 2021 vaccination uptake was associated with a 0.105% monthly increase in 2022 all-cause mortality [8]. A US nursing home study of 15,022 facilities found that elevated mortality after a COVID-19 infection persisted for 5 weeks in fully vaccinated residents but only 1 week in unvaccinated residents [10]. These ecological and observational studies cannot prove causation either, but they are consistent with a signal that the vaccines may carry a net mortality cost over time, especially with repeated boosting.

The known serious adverse events are real. Myocarditis after mRNA vaccines occurs at rates of 50-139 cases per million in male adolescents and 28-147 per million in young adult males [20]. A systematic review found that over half of patients with vaccine-related myocarditis still had persistent abnormalities at 3 months of follow-up [20]. Thrombosis with thrombocytopenia (VITT) after adenoviral vector vaccines is rare but often fatal, with an incidence of about 3.8 cases per million doses for the J&J vaccine and 16.3 per million for the AstraZeneca vaccine [43]. A Swedish nationwide analysis of 26 million doses found 10 deaths causally linked to vaccination out of 456 fatal reports, mostly from myocarditis, VITT, and vaccine failure [47].

The bottom line is that the evidence base is observational, conflicted, and incomplete. No randomized trial has ever compared vaccinated to unvaccinated people on all-cause mortality over multiple years. The studies that show benefit are vulnerable to healthy-vaccinee bias. The studies that show harm are vulnerable to confounding by frailty. The truth is probably somewhere in between: the vaccines likely reduce COVID-19 mortality in the short term for older adults and those with risk factors, but may increase all-cause mortality over longer time horizons, especially in younger people and with repeated boosting. The burden of proof has not been met for a blanket "safe" claim.

My call: the COVID-19 vaccines are not proven safe for all populations and all time frames. The evidence supports net benefit for older adults in the first year of vaccination, but the data also shows signals of net harm in younger people and with repeated doses. Confidence: moderate that the balance varies by age, dose number, and time since vaccination.