On the Friday after Thanksgiving 2025, Vinay Prasad, MD, MPH, who was then the director of the Food and Drug Administration's (FDA’s) Center for Biologics Evaluation and Research (CBER) sent a memo to staff. "For the first time," Prasad wrote, "the US FDA will acknowledge that COVID-19 vaccines have killed American children."
He said no fewer than 10 of 96 pediatric deaths reported to the federal Vaccine Adverse Event Reporting System (VAERS) from 2021 to 2024 were caused by COVID-19 vaccination. He said the coding was conservative. He said the actual number was higher.
The memo offered no clinical evidence. It accused the FDA of a years-long failure to acknowledge what new leadership had now uncovered.
The underlying analysis has now been revealed. It is a 73-page consensus review conducted by the career scientists in CBER's Office of Biostatistics and Pharmacovigilance (OBPV), the office responsible for post-market vaccine safety surveillance. It is dated December 5, 2025. It was released last week, attached as an exhibit to a letter Sen Ron Johnson (R-Wisc) sent to Department of Health and Human Services Secretary Robert F. Kennedy Jr.
The report does not support what the Prasad memo said.
A real risk, honestly characterized
Myocarditis after messenger RNA (mRNA) COVID-19 vaccination is a known and labeled risk. It has been on the prescribing information for Comirnaty (the Pfizer-BioNTech vaccine) and Spikevax (Moderna) since 2021. Fatal outcomes have not been described. The OBPV team identified that gap and recommended fixing it.
The multicenter MACiV study of 333 young patients with vaccine-associated myocarditis at 38 US hospitals reported zero cardiac deaths and zero heart transplants, though late gadolinium enhancement persisted in 60% of patients who had follow-up cardiac magnetic resonance imaging (MRI).
The international evidence points in the same direction. Copland and colleagues linked national data on 5.1 million children in England and found a small excess of myocarditis in adolescents after the second Pfizer-BioNTech dose, and none in 5- to 11-year-olds.
Canadian active surveillance found that vaccine-proximate myocarditis and pericarditis cases required intensive care in less than 7% of cases, compared with 28% among unvaccinated cases. A Korean national surveillance study of 3.7 million adolescents who received 8.1 million Pfizer-BioNTech doses identified 184 Brighton-confirmed myocarditis or pericarditis cases; 89.1% were mild, 20 were severe, two required extracorporeal membrane oxygenation, but no deaths were reported.
A French nationwide cohort of 4,635 hospitalized myocarditis patients aged 12 to 49 found no in-hospital deaths among the 558 post-vaccine cases, and a death rate of 0.2% over the next 18 months, compared with 1.3% for myocarditis after COVID-19 infection.
The labeling change reflects that picture honestly. The Prasad memo does not.
Fatal myocarditis after mRNA COVID-19 vaccination is rare but real. Pediatric cases have been reported, including in the two adolescent boys described by Gill and colleagues after a second Pfizer-BioNTech dose. The signal is well-characterized: rare, usually mild, sometimes severe, and occasionally fatal.
The labeling change reflects that picture honestly. The Prasad memo does not.
No deaths definitively tied to COVID vaccines
The OBPV team reviewed every pediatric death reported to VAERS after a COVID-19 vaccine from 2021 to 2024. They reviewed available medical records, autopsy reports, death certificates, and follow-up information collected through VAERS. They applied the World Health Organization (WHO)–Uppsala Monitoring Centre (UMC) causality framework, the standard tool in pharmacovigilance, with a pediatric cardiologist consulting on every cardiac case.
Of 96 deaths, they classified zero as certain. Two as probable. Five as possible. Sixty-two as unlikely. Twenty-seven as unassessable.
The framework's vocabulary matters. "Possible" means an event has a reasonable temporal relationship to vaccination but could also be explained by disease or other drugs. "Probable" means an alternative explanation is unlikely but cannot be ruled out. Neither category establishes causation. The report says this plainly: causality assessments "cannot definitively prove the connection between a drug or vaccine and a given adverse event."
The seven possible and probable cases all involved cardiac events. Four involved myocarditis. One involved myocarditis with stress cardiomyopathy. One involved cardiomyopathy without myocarditis. One involved cardiac arrhythmia. The distribution context is more than 138 million US doses of pediatric-labeled Comirnaty and Spikevax products. The report properly notes that distribution data cannot serve as true denominator data.
The OBPV team's recommendation is calibrated to that picture. They recommended a class safety labeling change under section 505(o)(4) of the Federal Food, Drug, and Cosmetic Act: a Boxed Warning, updated Warnings and Precautions, updated Adverse Reactions, and an updated Patient Package Insert.
The report states explicitly that it "does not constitute a benefit-risk analysis." It recommends no changes to who should receive the vaccine.
The report states explicitly that it 'does not constitute a benefit-risk analysis.'
This is what a post-market safety surveillance system is supposed to do. It identifies rare serious outcomes without pretending case reports can answer questions they cannot answer.
What “probable” and “possible” look like
Consider the cases that anchor the review.
The first involves a 7-year-old boy whose case is classified as probable. He had no history of noteworthy medical issues. Twelve days after his first Pfizer-BioNTech dose he developed fever, abdominal pain, and lethargy, and died of cardiac arrest two days later.
The autopsy showed lymphocytic myocarditis. Pathology testing by the Centers for Disease Control and Prevention (CDC) confirmed extensive heart muscle injury and found no molecular evidence of SARS-CoV-2, enteroviruses, parvovirus B19, influenza, cytomegalovirus, or adenoviruses. No specific alternative explanation was identified. Twelve days is within the known mRNA myocarditis risk window.
The reviewers called it probable. This is what "probable" looks like.
One of the possible cases involves a 13-year-old girl with elevated body mass index, asthma, and scoliosis. Fifteen days after her first Pfizer-BioNTech dose she developed shortness of breath, dizziness, and vomiting, and died the following day. The autopsy diagnosis was fulminant pancarditis with extensive heart muscle cell death. The autopsy report itself stated that "a direct cause and effect relationship between the vaccine and the development of myocarditis in this case cannot be determined without first excluding other potential causes of myocarditis."
CDC pathology then identified parvovirus B19, which can cause severe or fatal myocarditis that can mimic ischemic heart disease. The reviewers held the case at possible, because the temporal association was real and the alternative was real.
This is what "possible" looks like.
The remaining cases follow the same logic. The other probable case involved a 13-year-old boy with myocarditis on autopsy. CDC pathology also identified Clostridium septicum and concluded the death was attributable to clostridial sepsis, but OBPV reviewers noted the bacterial findings may have represented postmortem overgrowth, and the authors of the Gill paper argued they were a postmortem artifact rather than active sepsis.
The four other possible cases all involved a documented competing explanation, including parvovirus B19, uncontrolled type 1 diabetes, vasculitic concerns, or chronic structural heart disease. The categories were doing exactly what they were designed to do.
The 3 misleading moves
Prasad's memo relies on three specific misrepresentations.
First, it converts the WHO-UMC categories into causation. They are not built to support that move. A "possible" classification is a statement that the evidence is consistent with more than one explanation. It does not attribute vaccination to the death.
Second, the memo says the coding was "conservative" in the direction of exonerating the vaccine, and that vaccines are "exculpated rather than indicted in cases of ambiguity." The reviewer comments show the opposite. The OBPV team kept five cases in the "possible" category despite documented alternatives. That is what "possible" is for. If anything, the agency erred toward inclusion in the ambiguous cases.
Third, the memo places the findings in a political frame the analysis does not contain. It describes children "coerced, at the behest of the Biden administration." It concludes that vaccine regulation "may have harmed more children than we saved." The report declined to make any benefit-risk claim, by explicit design.
The memo places the findings in a political frame the analysis does not contain.
These are not small distinctions. They are the difference between a safety labeling change and a manufactured scandal.
What the documents actually say
Fatal vaccine-associated myocarditis is real. The international literature has documented it. The OBPV report's two probable pediatric cases are consistent with that literature. The five possible cases are exactly what the WHO-UMC framework was designed to capture: a temporal association, a documented alternative explanation, and not enough evidence to decide.
What the report does not show is that the FDA had been concealing pediatric vaccine deaths, that 10 US children died because of vaccination, or that the risk-benefit balance of pediatric COVID-19 vaccination has changed. The career scientists who wrote the report made none of those claims. The official who commissioned it made all of them.
The recommended labeling change is the appropriate response to what the report found. The misrepresentation of what it found is the problem.
Both documents are public. The difference is the story.
