Japan Vaccine-Death Records and the Viral Instagram Claim

Managing Expectations Health · June 22, 2026 · Japan / COVID vaccine records / excess mortality / Zenodo preprint / Instagram source check

The key finding is this: the Reel is pointing to a real Zenodo preprint involving Japanese disclosure-request data and Dr. Yasufumi Murakami. But the Instagram caption overstates the source. The record is a preprint, not a peer-reviewed verdict. It analyzes about 17.5 million administered doses among about 4.0 million individuals — not “18 million people.” Most importantly, the paper itself says it could not obtain the mortality data needed for a standard vaccinated-versus-unvaccinated comparison.

Medical caution

This article is source review, not medical advice. Do not use a social-media “vaccine detox protocol” as treatment, diagnosis, or substitute for qualified medical care. If someone has symptoms after vaccination or infection — chest pain, shortness of breath, fainting, neurologic symptoms, new severe fatigue, clot symptoms, or heart concerns — seek medical assessment.

Japan vaccine records Instagram claim source-check card

The Instagram claim

The Reel from mommavspharma claims:

Japan released a “bombshell study” comparing deaths of “18 million people,” vaccinated versus unvaccinated.
Dr. Murakami warned: “The more doses you get, the sooner you’re likely to die.”
No noticeable death spike appeared in the unvaccinated.
A peak appeared among vaccinated people, especially 90 to 120 days after the shot.
A graph allegedly showed that as dose count increased, the peak in deaths appeared sooner.
The post ends with a call to “Comment GOLD” for a “vaccine detox protocol.”

The audio in the Reel is not a narration of the Japanese data. It is an older clip of a masked healthcare worker at a CHI Memorial backdrop. The study claim is in the caption/overlay, not in the audio.

The source located

The likely underlying source is a Zenodo record published February 16, 2026:

Citizen volunteers’ disclosure request and web database for post-vaccination deaths among 4 million mRNA COVID-19 vaccine recipients in Japan: A massive wave of deaths occurring months after vaccination, amplified and prolonged by frequent dosing.

Zenodo lists it as a preprint. The creators include Hiroshi Arakawa, Takayasu Hamao, the Disclosure Request Team of United Citizens for Stopping mRNA Vaccines and Yukoku Union, and Yasufumi Murakami. The DOI is 10.5281/zenodo.18649880.

What the preprint says

Item	What the source says	Why it matters
Dataset scale	17,545,662 doses administered to 4,025,948 individuals used for statistical analysis.	The Reel’s “18 million people” wording is wrong or misleading. It is closer to 17.5 million doses, not 18 million individuals.
Data source	Citizen volunteers submitted disclosure requests to municipalities. 57 municipalities disclosed data enabling analysis.	This is not a full national registry analysis with complete individual-level covariates.
Claimed signal	The authors report mortality waves months after vaccination and dose-related timing patterns.	This is the real claim behind the Reel.
Estimate	The paper estimates about 3.89 million people died within one year of vaccination in Japan between 2021 and 2024.	The paper explicitly says this does not necessarily mean all were directly caused by vaccination.
Unvaccinated comparison	The authors say they could not obtain mortality data for the unvaccinated population itself.	This directly weakens the Reel’s “vaccinated vs. unvaccinated” framing.
Data quality	The authors acknowledge varying municipal data quality, inconsistent reporting standards, and incomplete information.	Those are major limitations for causal claims.

The most important limitation

The paper’s discussion says that, as a limitation of the disclosure request, the authors were unable to obtain mortality data for the unvaccinated population itself, and could not obtain data comparing healthy and unhealthy individuals. It says this made it difficult to use standard epidemiological methods comparing vaccinated and unvaccinated individuals, or healthy and unhealthy individuals.

That sentence is the center of the source-check. It means the Reel’s clean “vaccinated versus unvaccinated” claim goes beyond the study’s own stated ability.

Why timing curves do not prove cause by themselves

A graph showing “days from final vaccination to death” can be useful for hypothesis generation. But it is not enough to prove cause. To make a strong causal claim, an analysis needs careful handling of:

Age. Seniors were more likely to receive multiple doses and also have much higher baseline mortality.
Health status. Frail people, high-risk people, healthcare workers, and people with underlying disease may have different vaccination patterns.
Calendar time. Dose campaigns overlapped with COVID waves, seasonal mortality, healthcare pressure and Omicron-period effects.
Person-time denominators. Proper mortality rates require tracking who was at risk during each time period, not just counting deaths after a final dose.
Healthy vaccinee / unhealthy stopper effects. People healthy enough to keep receiving boosters can differ from people who stop after a dose, and those differences can distort curves.
Cause of death. All-cause death timing does not by itself identify vaccine causality, COVID infection, cancer, heart disease, aging or other causes.

Peer-reviewed context

The Zenodo preprint should be read alongside peer-reviewed Japanese safety work, not in isolation.

A 2022 paper in Vaccine used administrative claims data linked with a vaccination registry in a Japanese city and evaluated serious outcomes and all-cause mortality 21 days after mRNA vaccination. Its conclusion: the vaccine was generally safe, while noting a pulmonary-embolism signal after first dose in women.
A 2023 Cureus paper analyzed reported deaths after BNT162b2 using sex ratios and short risk windows. It suggested vaccination may influence reported deaths during a short risk period, but also said a Japanese cohort study found no significant increase in all-cause mortality owing to vaccination and that such a finding supports vaccine safety. It did not claim that all excess deaths were vaccine deaths.
Japan’s excess-deaths dashboard shows observed and expected deaths, but excess mortality dashboards do not by themselves assign cause.

What is fair to say

Fair: the Reel points to a real Zenodo preprint and public database project.
Fair: the preprint makes serious claims about delayed mortality patterns after vaccination and says the data deserve attention.
Fair: Japan’s excess mortality and vaccine-safety data transparency are legitimate public-interest topics.
Not fair: saying the source cleanly compared 18 million vaccinated versus unvaccinated people.
Not fair: treating a time-since-final-dose death curve as proof that “the more doses you get, the sooner you die.”
Not fair: using the claim as a funnel for a “vaccine detox protocol.”

Managing expectations

The Murakami-linked Zenodo record is not nothing. It is a real disclosure-request project and a real preprint making a strong mortality-signal argument. It should be examined, replicated and critiqued openly.

But the Instagram post turns a limited, controversial preprint into a much stronger public claim. It changes doses into people, implies a clean vaccinated-versus-unvaccinated comparison, and moves from data discussion into detox-protocol marketing. That is exactly where Managing Expectations should slow the story down.

The bottom line: real source, overstated Reel. Worth investigating, not proof by itself.

Source links

Health source library

Back to Managing Expectations Health for source-grounded reviews of medical claims, doctor interviews, papers and viral narratives.

Back to Health