Debunks › Vaccines & Health › Anti-Vaccine Claims

Testing Anti-Vaccine Claims

Vaccines are neither magical shields nor uniquely dangerous poisons. They are medical products, and medical products can be tested. The useful question is not whether one is “pro-vaccine” or “anti-vaccine,” but whether a specific claim about a specific vaccine, outcome, population, dose, or ingredient survives contact with the evidence. Some vaccine claims fail spectacularly. Some objections are based on genuine but badly interpreted safety signals. Some risks are real. Some policy disputes cannot be settled by laboratory data at all.

This article takes the claims seriously enough to separate them. It does not ask the reader to trust a government agency, pharmaceutical company, physician, influencer, or slogan. Instead it asks what each hypothesis predicts, what kind of evidence could confirm or falsify it, what the strongest available studies actually show, and whether the same standard is being applied to vaccination and to the infections vaccination is intended to prevent.

Contents

• 1. The standard: what would count as evidence?
• 2. “Vaccines do not work”
• 3. “Sanitation, not vaccines, caused the decline in disease”
• 4. “Natural immunity is always better”
• 5. “Vaccines cause autism”
• 6. The Wakefield/MMR story
• 7. “Mercury in vaccines causes neurological damage”
• 8. Aluminum adjuvants
• 9. Formaldehyde, polysorbate, antibiotics, and the word “toxin”
• 10. “Too many, too soon”: immune-system overload
• 11. “The childhood schedule was never tested”
• 12. “Vaccines are never tested against saline placebo”
• 13. “There are no long-term safety studies”
• 14. VAERS: signal detection is not causation
• 15. SIDS and infant deaths
• 16. Autoimmune and chronic disease claims
• 17. Fertility, pregnancy, and HPV claims
• 18. mRNA, DNA, “gene therapy,” and shedding
• 19. Myocarditis: a real adverse event and why nuance matters
• 20. Breakthrough infections do not mean zero effectiveness
• 21. “The flu shot barely works”
• 22. “Why vaccinate a newborn against hepatitis B?”
• 23. HPV vaccination and cancer prevention
• 24. Vaccine shedding and vaccine-derived polio: the real exception
• 25. Real vaccine injuries and historical safety failures
• 26. “Unvaccinated children are healthier”
• 27. “The package insert lists death/autism/etc., so the vaccine causes it”
• 28. Fetal cell lines: what is actually in the vial?
• 29. SV40-contaminated polio vaccine and cancer claims
• 30. Antibody-dependent enhancement and Dengvaxia: a real warning
• 31. “Vaccine manufacturers have no liability”
• 32. Pharmaceutical profit, regulators, and conflicts of interest
• 33. Mandates are a policy question, not a biological result
• 34. Claim-by-claim verdicts
• Conclusion

1. The Standard: What Would Count as Evidence?

A vaccine-safety claim is a causal claim. If someone says “vaccine X causes condition Y,” the fact that Y happened after X is only the beginning of the investigation. Human beings develop seizures, strokes, miscarriages, autoimmune disease, cancer, autism diagnoses, heart attacks, and sudden deaths whether or not they were recently vaccinated. When millions of people receive an intervention, thousands of unrelated medical events will occur afterward by coincidence. The question is whether the event occurs more often than its expected background rate, whether the timing fits a biologically plausible mechanism, whether the association repeats in independent data, whether risk changes with exposure, and whether alternative explanations can be excluded.

This is why anecdotes matter as signals but are weak as estimates of causation. “My child changed after a vaccine” may be a completely sincere observation and still fail to distinguish coincidence from cause. Conversely, “most people were fine” cannot disprove a rare one-in-50,000 adverse event. Detecting common side effects can be done in ordinary clinical trials. Detecting very rare harms usually requires surveillance across hundreds of thousands or millions of doses after authorization. A competent system therefore needs both pre-licensure trials and post-market pharmacovigilance.

The same discipline must be applied to benefits. A vaccine that reduces infection by 60% is not 100% effective, but it is also not “useless.” A vaccine that barely changes infection risk in one season might still reduce hospitalization. Protection may be strong against one variant and weaker against another. Immunity can wane. Risk can differ by age and sex. A benefit-risk assessment for a 75-year-old need not be identical to one for a healthy 20-year-old. The evidence should be allowed to stay complicated when reality is complicated.

2. “Vaccines Do Not Work”

The broad claim that vaccines do not work is one of the easiest to test because infectious diseases leave population-level fingerprints. Before measles vaccination began in the United States in 1963, CDC estimates that roughly 3 to 4 million Americans contracted measles each year. About 500,000 cases were reported; among those reported cases, roughly 400 to 500 people died, 48,000 were hospitalized, and about 1,000 developed encephalitis.^[1] Two doses of modern measles vaccine are about 97% effective against measles, while one dose is about 93% effective.^[2]

Pertussis offers another natural experiment. Before widespread vaccination in the 1940s, the United States reported more than 200,000 cases annually. After widespread childhood vaccination began, annual incidence fell by more than 90% compared with the pre-vaccine era.^[3] Pertussis later resurged for several reasons—including waning immunity and limitations of the acellular vaccine—but a resurgence from a very low post-vaccine baseline is not evidence that the original decline never occurred.

Polio is even harder to explain away. The United States recorded more than 20,000 cases of permanent paralytic poliomyelitis in 1952. Effective vaccines arrived in 1955 and the early 1960s. By 1965 fewer than 100 cases were being reported, and by 1973 fewer than 10.^[4] This timing is not a slow century-long trend in plumbing. It is a sharp epidemiological change coinciding with the intervention, replicated in country after country as vaccine coverage rose.

Chickenpox supplies a modern example in an era when American sanitation was already excellent. During the first 25 years of the U.S. varicella vaccination program, hospitalizations among people under 50 fell about 94% and deaths about 97%. Among people under 20—those born into the vaccination era—hospitalizations fell about 97% and deaths by more than 99%.^[5] It is difficult to attribute a 1995-onward collapse in severe chickenpox to a sudden revolution in sewer construction.

At the global scale, a 2024 WHO-led analysis estimated that immunization against 14 diseases had saved at least 154 million lives over the previous 50 years, with infants accounting for most of the lives saved.^[6] Such estimates depend on models and assumptions and should not be treated as divinely revealed exact integers. But the underlying record—eradication of smallpox, near-eradication of wild polio, collapse of Hib disease where conjugate vaccine is introduced, and repeated rebounds when coverage falls—does not support the proposition that vaccination has no biological effect.

Verdict: False as a general claim. Vaccine effectiveness varies by vaccine, outcome, strain, age, and time since vaccination, but multiple vaccines produce large, independently measurable reductions in disease.

3. “Sanitation, Not Vaccines, Caused the Decline in Disease”

This argument begins with a truth and then overextends it. Clean water, sewage disposal, nutrition, less crowding, vector control, antibiotics, intensive care, and improved living conditions dramatically reduced infectious-disease mortality. A graph of deaths from measles or scarlet fever can therefore fall long before a vaccine appears. But mortality is not the same variable as incidence. Better supportive care can make an infection less likely to kill you without preventing the infection from occurring.

Measles illustrates the distinction. By the early 1960s, Americans were far less likely to die from measles than they had been a century earlier, but millions were still infected annually. After vaccination began, cases collapsed. The sanitation-only hypothesis has to explain why an intervention that should mainly affect waterborne or fecal-oral transmission suddenly synchronized with a collapse of a highly contagious airborne virus. It also has to explain why measles returns in under-vaccinated communities that possess modern toilets, treated water, refrigeration, and hospitals.

The same logic applies to varicella. American sanitation in 1995 was not primitive. Yet after varicella vaccine introduction, cases, hospitalizations, and deaths fell dramatically.^[5] Smallpox eradication is even less compatible with the claim: improved sanitation did not merely reduce smallpox mortality; coordinated vaccination and surveillance eliminated natural transmission worldwide. The last naturally acquired case occurred in 1977, and eradication was declared in 1980.

Sanitation and vaccination are complementary interventions, not mutually exclusive explanations. If one improves roads and another improves brakes, evidence for safer roads does not prove brakes are useless. Historical mortality graphs are valuable, but the correct test is disease-specific and asks about incidence, severity, timing, coverage, biological mechanism, and what happens when vaccination is introduced or withdrawn.

Verdict: Misleading. Public-health improvements reduced infectious-disease mortality enormously, but they cannot account for the disease-specific collapses in incidence that followed vaccination.

4. “Natural Immunity Is Always Better”

Infection can generate powerful immunity. For some diseases, surviving infection produces durable protection. That fact does not establish that infection is the preferable route to immunity. The comparison is not “natural immunity versus no immunity.” It is normally “immunity acquired by suffering the disease” versus “immunity acquired through an intervention designed to expose the immune system to a safer representation of the pathogen.”

A child who survives measles may indeed emerge with measles immunity, but the price of acquiring it was measles—with its risks of pneumonia, encephalitis, hospitalization, death, and the less common late complication subacute sclerosing panencephalitis. A person who survives tetanus cannot even assume reliable protective immunity from the illness. A person who acquires hepatitis B in infancy faces a much greater risk of chronic infection than an adult who acquires it later. “Natural” describes origin; it does not rank outcomes.

Natural infection can also impose costs that vaccination does not reproduce. Some pathogens damage the very systems that respond to them. Others establish latency or chronic infection. Some have toxin-mediated pathology where very little toxin is sufficient to cause severe disease. In such cases, saying “the disease gives better immunity” ignores the adverse event that matters most: the disease itself.

There are also circumstances where prior infection changes the marginal benefit of vaccination. That is a legitimate scientific question and has been especially important in COVID-19 policy. It does not rescue the universal slogan. The correct answer depends on the pathogen, prior exposure, time since infection, age, immune status, circulating strain, and desired endpoint.

Verdict: Sometimes infection produces strong immunity, but “always better” is false because it ignores the morbidity required to obtain that immunity and the disease-specific differences in immune response.

5. “Vaccines Cause Autism”

This is the most famous vaccine-safety claim and one of the most extensively tested. It has been examined from several angles: MMR exposure, thimerosal exposure, total vaccine exposure, timing, genetically or clinically “susceptible” subgroups, and population trends after vaccine ingredients changed. A causal hypothesis should survive at least some of these tests. The autism-vaccine hypothesis has not.

A 2019 Danish nationwide cohort followed 657,461 children and compared autism incidence in MMR-vaccinated and unvaccinated children while adjusting for age, sex, birth year, other childhood vaccines, family history, and other risk factors. The adjusted hazard ratio was 0.93 with a 95% confidence interval from 0.85 to 1.02—no increased risk. The investigators also found no evidence that MMR triggered autism in higher-risk subgroups or produced a clustering of diagnoses after vaccination.^[7]

The result is not standing alone. Reviews and meta-analyses incorporating hundreds of thousands to more than a million children have repeatedly failed to detect the predicted relationship. In 2026 WHO published an updated evidence review covering peer-reviewed research from 2010 through August 2025. Its central conclusion was that the most methodologically rigorous studies did not support a causal association between vaccines—including thiomersal-containing vaccines—and autism spectrum disorder. Studies reporting positive associations were judged to have low evidentiary strength and high risk of bias, while the included meta-analyses did not support causation.^[8]

One common response is that epidemiological studies can miss a “susceptible subgroup.” That possibility is scientifically reasonable in principle, but it must generate a testable subgroup definition. If every negative study is answered by moving the alleged effect into an undefined population that cannot be prospectively identified, the claim becomes unfalsifiable. The Danish study specifically tested family history and other autism-risk strata and still did not find the hypothesized signal.^[7]

Another response is that autism prevalence has increased while vaccine use expanded. But prevalence trends alone cannot establish cause because diagnostic criteria, awareness, services, ascertainment, age at diagnosis, and classification all changed over the same period. More importantly, the causal prediction can be tested directly in individuals and in populations where formulations changed. Removing thimerosal from most U.S. childhood vaccines did not produce the predicted collapse in autism diagnoses. MMR never contained thimerosal in the first place.^[9]

Verdict: The claim is not supported. The vaccine-autism hypothesis has been tested repeatedly with large datasets and multiple designs, and the predicted association has not appeared.

6. The Wakefield/MMR Story

The modern MMR-autism panic is inseparable from a 1998 Lancet paper led by Andrew Wakefield. The paper described 12 children with gastrointestinal symptoms and developmental disorders and suggested a possible new syndrome linked temporally to MMR vaccination. Twelve children in an uncontrolled case series could never establish population-level causation, even if every record had been pristine. The paper nevertheless became a media catalyst.

Subsequent investigation exposed serious problems with the work, and The Lancet ultimately retracted it. The episode is sometimes described too simply as “one paper was wrong.” The deeper methodological lesson is that the claim failed replication. If MMR were a substantial cause of autism, large controlled studies should show an increased rate among vaccinated children, a temporal concentration after vaccination, or a dose/timing pattern. They do not.

It is also important not to use Wakefield as a substitute for answering every concern. Even a disgraced paper can accidentally point toward a true phenomenon. The decisive case against MMR-autism is therefore not “Wakefield was bad.” It is that the hypothesis was subsequently tested at scales many orders of magnitude larger and did not behave like a causal relationship.

Verdict: The original case series was incapable of establishing causation and was retracted; more importantly, the proposed association failed large-scale replication.

7. “Mercury in Vaccines Causes Neurological Damage”

Thimerosal is an organomercury-containing preservative historically used in some multidose vaccine vials to prevent bacterial and fungal contamination. The word “mercury” understandably alarms people because some mercury compounds are neurotoxic. But toxicology depends on chemical species, dose, route, and pharmacokinetics. Ethylmercury derived from thimerosal is not pharmacokinetically identical to methylmercury, the form central to many dietary-exposure warnings.

More importantly, the autism hypothesis generated a straightforward natural experiment. Thimerosal was removed from or reduced to trace amounts in vaccines routinely recommended for young children in the U.S. beginning around 2001; MMR, varicella, inactivated polio, and pneumococcal conjugate vaccine did not contain it as a preservative. Some multidose influenza formulations have continued to use it, while thimerosal-free alternatives exist.^[9] If thimerosal were a major driver of autism prevalence, a substantial drop should have followed its removal from routine pediatric exposure. That did not happen.

The updated WHO review through 2025 specifically evaluated thiomersal-containing vaccines and autism and found that the strongest studies and meta-analyses did not support causation.^[8] This does not imply that “mercury is safe at any dose.” That would be equally unscientific. It means the exposure and outcome actually under dispute have been studied, and the predicted neurological signal has not been demonstrated.

Verdict: The presence of an organomercury preservative is not evidence of poisoning, and large-scale evidence does not support thimerosal as a cause of autism.

8. Aluminum Adjuvants

Aluminum salts are used in some non-live vaccines as adjuvants: ingredients that increase the immune response to an antigen. This allows a vaccine to produce stronger or more durable immunity with a practical antigen dose. Aluminum adjuvants can increase local reactogenicity—pain, redness, swelling—and they are not present in every vaccine. FDA and CDC describe decades of use and note that aluminum is also ubiquitous in food, water, and the environment.^[10][11]

A common argument compares injected aluminum with an oral exposure limit and concludes that any injected amount above that number must be toxic. That comparison can be misleading in both directions. Injection and ingestion have different absorption kinetics, so one should not casually convert between them. But the correct response is not to declare injected aluminum harmless by definition; it is to study actual vaccine exposures and actual health outcomes.

A 2025 Danish nationwide cohort examined 1,224,176 children and used variation in cumulative aluminum exposure from early-childhood vaccines to test associations with 50 autoimmune, allergic/atopic, and neurodevelopmental outcomes. The published analysis found no increased rate of any of the grouped outcome classes with greater aluminum exposure and reported that moderate-to-large risk increases were inconsistent with the data for most outcomes.^[12]

That paper has since attracted methodological criticism and a 2026 reanalysis arguing that some neurodevelopmental results deserve a different interpretation. That debate is worth acknowledging rather than pretending the literature froze in 2025. It is also why one should not rest the entire safety case for aluminum on a single cohort. WHO's vaccine-safety committee reviewed aluminum evidence alongside the broader autism literature and concluded that available high-quality evidence did not support aluminum-containing vaccines as a cause of autism.^[13] The reasonable position is therefore neither “aluminum can never cause harm” nor “aluminum is a metal, therefore vaccines containing it are poisonous.” It is that exposure-specific safety questions require outcome data, and the current body of evidence has not established the sweeping chronic-disease claims commonly made online.

Verdict: Aluminum adjuvants are biologically active and can contribute to local reactions, but the claim that routine vaccine aluminum has been shown to cause autism or broad chronic disease is not supported by the weight of current evidence. Ongoing study is appropriate.

9. Formaldehyde, Polysorbate, Antibiotics, and the Word “Toxin”

Ingredient lists are among the most persuasive anti-vaccine images because chemistry names look frightening when stripped of dose and purpose. Formaldehyde can be hazardous. Aluminum can be toxic at sufficient exposure. Some antibiotics can cause allergic reactions. Even sodium chloride can be lethal in sufficient quantity. Toxicology therefore cannot stop at “this chemical can be dangerous.” The scientifically meaningful sentence requires a dose, route, duration, and endpoint.

Formaldehyde is used during manufacture of some vaccines to inactivate viruses or detoxify bacterial toxins. Residual amounts can remain after purification. FDA notes that formaldehyde is also a normal product of human metabolism and that the amount naturally present in the body of a newborn is dozens of times greater than the upper amount received from a vaccine dose or vaccines given over time.^[10] This does not mean inhaling industrial formaldehyde is safe. It means hazard and exposure are different questions.

Trace antibiotics may be used during manufacture to suppress bacterial contamination; purification reduces them to very small or undetectable quantities. Preservatives prevent contamination of multidose containers. Sugars, amino acids, gelatin, and proteins may stabilize a product during storage or freeze-drying. Surfactants can keep components mixed. An ingredient's unfamiliar name does not establish that the final concentration is dangerous.

Allergy is a legitimate exception to casual reassurance. A person with a severe allergy to a vaccine component may have a contraindication or need specialist guidance. Rare anaphylaxis is a known potential adverse event after vaccination, which is why vaccination sites are equipped to recognize and treat it. The existence of rare allergy, however, is very different from the assertion that every ingredient is systemically “toxic.”

Verdict: Ingredient-list arguments usually confuse hazard with dose and route. Some components can cause reactions in susceptible people, but the mere presence of a chemical with a frightening safety data sheet at industrial exposure does not demonstrate toxicity at vaccine exposure.

10. “Too Many, Too Soon”: Immune-System Overload

The immune system does not count needles. It recognizes molecular structures—antigens—and coordinates innate and adaptive responses to them. Infants begin encountering enormous numbers of antigens from skin, food, airways, gut microbes, and ordinary infections immediately after birth. A schedule containing several injections can look visually dramatic while representing a tiny fraction of the antigenic challenges handled every day.

Modern vaccines can also contain fewer antigenic components even when they protect against more diseases because purified or recombinant technologies replace whole organisms with selected proteins or polysaccharides. The number of injections is therefore a poor proxy for immunological burden.

The overload hypothesis can be tested by asking whether greater cumulative vaccine-antigen exposure predicts later neuropsychological impairment. A 2013 study evaluated 1,047 children using standardized measures of intelligence, language, verbal memory, attention, executive function, tics, achievement, visual-spatial ability, and behavioral regulation. It found no adverse association between antigen exposure during the first two years of life and the later neuropsychological outcomes assessed.^[14]

That does not prove that every conceivable schedule modification is equivalent. Timing is chosen for epidemiological and immunological reasons: when maternal antibodies wane, when infants become vulnerable, when a product has demonstrated immunogenicity, and when boosting is needed. Specific vaccines can interfere or interact, which is why schedules and coadministration are studied. “The immune system can handle many antigens” is not a license to inject arbitrary substances. It is simply why the raw number of shots does not establish overload.

Verdict: The visual impression of “many shots” is not an immunological metric, and studies of cumulative antigen exposure have not supported the predicted developmental harm.

11. “The Childhood Schedule Was Never Tested”

This claim usually combines several different demands. Individual vaccines are tested before licensure. Combinations and coadministration are studied. Safety is monitored after licensure. But there has not been one single randomized trial assigning newborns to “the entire recommended schedule for years” versus “no vaccines at all” and then withholding all established vaccines from the control group. That statement is true. It does not follow that the schedule is therefore unstudied.

The U.S. National Academies examined this exact issue in a 2013 consensus report. The committee found no evidence of major safety concerns associated with adherence to the childhood immunization schedule, while also recommending better methods and continued research capable of detecting schedule-level effects.^[15] That is a more defensible conclusion than either extreme. “We have no evidence of major schedule-level harm” is not the same as “every possible long-term effect has been proven impossible.”

Medical evidence is often assembled from overlapping layers rather than one omnipotent trial: randomized studies of products, immunogenicity studies, coadministration trials, adverse-event comparisons, disease registries, claims databases, active surveillance networks, case-control studies, pregnancy registries, self-controlled case series, and natural experiments when recommendations change. Each design has weaknesses; convergence across different weaknesses is precisely what makes a causal conclusion stronger.

There is also an ethical problem with the proposed definitive experiment once effective vaccines exist. Randomly assigning a large cohort of children to receive no protection against measles, polio, pertussis, Hib, tetanus, and other established risks for years would knowingly withhold interventions already supported by evidence. Ethical constraints are not proof of safety, but they explain why “no whole-schedule versus zero-vaccine RCT” is not the devastating indictment it is presented as.

Verdict: Misleading. There is no single RCT of the entire childhood schedule versus lifelong nonvaccination, but the schedule is studied through multiple clinical and epidemiological designs, and major reviews have not found evidence of a broad safety problem.

12. “Vaccines Are Never Tested Against Saline Placebo”

Some vaccine trials use inert placebos; others use active comparators. FDA describes phase 2 and phase 3 vaccine trials as controlled studies in which the comparison group may receive an approved vaccine, a placebo, or another control substance.^[16] Therefore “vaccines are never placebo tested” is simply false. The stronger version—“not every vaccine trial uses saline”—is true.

Why use an active comparator? If an effective vaccine already exists, assigning participants to saline can expose them to preventable disease merely to recreate an experiment whose baseline efficacy has already been established. A new formulation may instead be compared with the licensed predecessor to determine whether protection and adverse-event rates are similar or improved. In other contexts, especially when no effective vaccine exists, placebo controls may be appropriate.

Another source of confusion is the word placebo. In a trial designed to blind participants to injection reactions, a control may contain a formulation buffer or another vaccine rather than pure saline. That choice can affect which question the trial answers. A study comparing new vaccine A with established vaccine B is excellent for asking whether A differs from B; it is less direct for asking whether either differs from absolutely no injection. The correct criticism should therefore identify the trial, comparator, and endpoint rather than declaring all vaccine trials fraudulent by category.

Rare adverse events are also usually beyond the statistical reach of pre-licensure trials regardless of comparator. A 20,000-person randomized trial is unlikely to detect an event that occurs once per 100,000 doses. That is why post-market surveillance is not an optional afterthought but part of the safety architecture.

Verdict: False in its categorical form. Vaccine trials can and do use inert placebos, but active comparators are also scientifically and ethically appropriate depending on the question.

13. “There Are No Long-Term Safety Studies”

The phrase “long term” is often left undefined. Does it mean six months, five years, twenty years, or a full human lifespan? If the demand is that no vaccine can be considered usable until an entire generation dies of old age, then no newly developed medical product could ever satisfy it. If the claim is that delayed or chronic outcomes are not studied at all, that is false.

Some vaccine platforms have been used for decades, allowing very long observational follow-up. Large national health registries can compare exposed and unexposed people over years. Post-market systems can detect unexpected patterns after millions of doses. Pregnancy outcomes, neurologic outcomes, autoimmune diagnoses, and cancers can be studied through linked medical records. Long-term safety is therefore not one study design but a set of surveillance and epidemiological methods.

There is a genuine limitation worth preserving: observational studies are susceptible to confounding, coding differences, loss to follow-up, and changes in diagnosis. RCTs usually have cleaner causal inference but shorter follow-up and smaller populations. Safety science works by combining designs. A rare effect that appears in spontaneous reports can be checked in active surveillance. A signal in a cohort can be tested elsewhere. If a claim requires dismissing every large negative study as corrupted while treating every positive case report as decisive, it is no longer using a symmetric standard of evidence.

Verdict: False as a blanket statement. Long-term and post-market vaccine safety research exists, although no study design can prove the absence of every conceivable rare effect forever.

14. VAERS: Signal Detection Is Not Causation

The U.S. Vaccine Adverse Event Reporting System is deliberately permissive. Patients, parents, clinicians, manufacturers, and others can submit reports of health events that occurred after vaccination. This openness is useful because a safety system should be able to notice unusual patterns without requiring the reporter to prove causation first. It also means the raw database contains events that were caused by vaccination, events that might have been caused by vaccination, duplicate or incomplete reports, and events that would have happened anyway.

CDC states explicitly that a VAERS report does not mean a vaccine caused the event and that VAERS data alone generally cannot be used to calculate how often an adverse event occurs in the population.^[17] This is not a disclaimer invented to hide inconvenient data. It is the mathematical consequence of the system's design: there is no reliable denominator for many analyses, reporting probability changes with publicity, and the database does not supply a matched unvaccinated comparison group.

Suppose 10 million older adults are vaccinated in a short period. Some number will have heart attacks the following week because older adults have heart attacks. If every post-vaccination heart attack is counted as “a vaccine death,” the method guarantees a false causal signal even if the vaccine has zero effect. The correct analysis asks whether the observed rate exceeds the expected background rate after adjusting for age, sex, time, season, comorbidity, and other factors.

VAERS nevertheless has real value. Unexpected clustering can prompt focused investigation. A pattern may then be evaluated in systems with known denominators and linked medical records. This signal-to-verification pipeline is how pharmacovigilance is supposed to work. Treating VAERS as useless is wrong; treating every report as a confirmed injury is also wrong.

Verdict: VAERS is an early-warning system, not a database of adjudicated vaccine-caused injuries. Raw report counts cannot establish incidence or causation.

15. SIDS and Infant Deaths

Vaccination and sudden infant death syndrome overlap in time. Infants receive multiple vaccines during the first months of life, and SIDS also peaks in infancy. Temporal overlap can create extraordinarily powerful anecdotes: a healthy baby is vaccinated, dies days later, and the vaccination becomes the most salient recent event. That sequence deserves investigation, but it is not sufficient to establish cause.

CDC's review of the literature states that multiple studies and safety reviews have not found childhood vaccination to cause or increase SIDS risk.^[18] An informative historical observation is that U.S. SIDS rates declined dramatically after infant sleep-position campaigns in the 1990s and then stabilized even while the number of recommended immunizations increased. That pattern is difficult to reconcile with a large causal effect from vaccination.

The right counterfactual is essential: among a million infants of the same ages who were not vaccinated on a given day, how many would die unexpectedly over the next week? Without that background, “deaths after vaccination” is a numerator without context. This same principle applies to miscarriage after vaccination, cancer diagnoses, strokes, and many other events.

Verdict: Temporal coincidence is expected because vaccination and SIDS occur in the same age window; population studies have not supported vaccination as a cause of SIDS.

16. Autoimmune and Chronic Disease Claims

Vaccines deliberately stimulate immunity, so asking whether they can precipitate autoimmune phenomena is not irrational. Infections themselves are well-known triggers of immune activation, and rare immune-mediated adverse events can occur after medical interventions. The mistake is jumping from “biologically conceivable” to “therefore vaccines are causing the modern burden of autoimmune disease.”

For a broad chronic-disease hypothesis, researchers can compare diagnoses across exposure levels and populations. The 2025 Danish aluminum cohort, for example, examined autoimmune, allergic, and neurodevelopmental outcomes across more than 1.2 million children and did not find increased grouped rates with higher vaccine-aluminum exposure.^[12] Other specific claims—such as hepatitis B vaccination causing multiple sclerosis—have been investigated without establishing the proposed association.

At the same time, there are genuine immune-mediated vaccine adverse events. The 1976 swine-flu vaccine was associated with an increased risk of Guillain-Barré syndrome. For seasonal influenza vaccines, when an increased GBS risk has been observed, CDC estimates it at roughly one to two additional cases per million doses; influenza infection itself is associated with GBS as well.^[19] That is exactly what a mature position looks like: not “immune-mediated harms are impossible,” but “specific products can carry specific small risks that must be quantified.”

The broad claim also suffers from disease heterogeneity. Type 1 diabetes, rheumatoid arthritis, multiple sclerosis, inflammatory bowel disease, lupus, asthma, eczema, and autism do not share one cause merely because they are placed on an internet list labeled “chronic disease.” Combining dozens of outcomes after the fact makes chance associations easier to find and causal interpretation harder.

Verdict: Vaccination can rarely be associated with specific immune-mediated adverse events, but evidence does not support the sweeping claim that routine vaccination is a major general cause of autoimmune or chronic disease.

17. Fertility, Pregnancy, and HPV Claims

Fertility rumors became especially prominent during the COVID-19 vaccine rollout. Claims ranged from antibodies attacking syncytin to altered menstrual cycles being interpreted as infertility. A temporary change in cycle timing is not equivalent to loss of fertility, and a molecular resemblance is not enough to demonstrate clinically meaningful cross-reactivity.

A 2022 systematic review and meta-analysis covering 29 studies examined sperm parameters and pregnancy outcomes and found no evidence that COVID-19 vaccination impaired male or female fertility. Pregnancy rates in vaccinated and unvaccinated groups did not show the predicted reduction, and pooled sperm measures did not show significant deterioration after vaccination.^[20] Like all early-pandemic literature, these data had limitations, but the feared large fertility effect did not materialize.

The same type of rumor has followed HPV vaccination. CDC and FDA safety monitoring has not found evidence that HPV vaccination causes primary ovarian insufficiency or general reproductive failure.^[21] A claim of widespread infertility should be relatively easy to detect after hundreds of millions of doses because it predicts changes in conception, ovarian function, or assisted-reproduction outcomes large enough to appear across datasets.

Pregnancy deserves product-specific caution. Some live attenuated vaccines are contraindicated during pregnancy because of theoretical or known fetal risk, whereas other vaccines are specifically studied for use during pregnancy to protect the mother or infant. Therefore “vaccines are safe in pregnancy” is too broad, just as “vaccines cause miscarriage” is too broad. The product, gestational timing, and disease threat matter.

Verdict: Broad infertility claims have not been supported by reproductive-outcome data. Pregnancy recommendations are vaccine-specific, not a universal property of all vaccines.

18. mRNA, DNA, “Gene Therapy,” and Shedding

mRNA vaccines use messenger RNA as a temporary instruction for cells to produce an antigen that the immune system can learn to recognize. Calling this “gene therapy” can be rhetorically effective because both technologies involve nucleic acids, but it obscures a crucial distinction: conventional gene therapy aims to alter, replace, or regulate genetic material in a sustained way. An mRNA vaccine supplies transient cytoplasmic instructions and does not require insertion into the genome.

CDC's explanation of the platform notes that the vaccine mRNA does not enter the cell nucleus, where chromosomal DNA is stored, and does not interact with or alter that DNA.^[22] It is possible to construct unusual laboratory systems in which RNA-derived sequences are reverse-transcribed under artificial conditions; that is not evidence that vaccination routinely integrates genetic material into the chromosomes of vaccinated humans. A clinically meaningful integration claim would require demonstration in human tissues at relevant frequency plus evidence of resulting pathology.

The word “genetic” also does not mean “changes your genes.” Every RNA virus carries genetic material. Protein production from messenger RNA is a normal cellular process. The relevant safety questions are about dose, biodistribution, inflammatory response, antigen expression, and adverse outcomes—not whether the molecule belongs to the broad category of genetic information.

COVID-19 mRNA vaccines also do not contain live replicating SARS-CoV-2, so they cannot “shed” infectious vaccine virus in the way a live attenuated vaccine can. CDC's archived myth guidance correctly distinguished shedding as a phenomenon that can occur with some live vaccines from the non-live mRNA platform.^[23] Claims that vaccinated people emit enough vaccine material or spike protein to alter the menstrual cycles, fertility, or health of nearby unvaccinated people have not been supported by exposure measurements or epidemiology.

Verdict: mRNA vaccination uses genetic instructions but has not been shown to rewrite the recipient's genome. Non-live mRNA vaccines do not shed infectious vaccine virus.

19. Myocarditis: A Real Adverse Event and Why Nuance Matters

Myocarditis after mRNA COVID-19 vaccination is a useful test of whether someone is reasoning about vaccine safety or defending a tribe. The association is real. It is concentrated especially in adolescent and young adult males and has been most evident after certain second mRNA doses. Denying that signal would be as unscientific as exaggerating it.

A large English analysis found that, averaged across the whole studied population, SARS-CoV-2 infection was associated with a larger excess of myocarditis than vaccination. But age, sex, and product changed the comparison. In men younger than 40, the study estimated about 97 excess myocarditis events per million after a second dose of Moderna's mRNA-1273 vaccine versus about 16 excess events per million after a positive SARS-CoV-2 test in that subgroup.^[24] A Nordic cohort likewise found the highest post-vaccination myocarditis risk in young males after the second mRNA dose, with a larger excess after mRNA-1273 than BNT162b2.^[25]

This is exactly why slogans fail. “COVID infection has more myocarditis than vaccination” can be true in an overall population while being false for a particular product/dose/age/sex comparison. Conversely, “the vaccine causes myocarditis” does not tell us whether the absolute risk is 1 in 100, 1 in 10,000, or 1 in a million, nor whether another vaccine, dose interval, or infection poses a larger risk.

Public-health recommendations changed as this evidence accumulated: dose intervals, product preferences, and age-specific guidance were adjusted in various countries. That adaptive behavior is what safety monitoring is for. Finding an adverse effect is not proof that surveillance failed; refusing to acknowledge one after it is found would be.

Verdict: True in a defined form: mRNA COVID-19 vaccines can rarely cause myocarditis, particularly in younger males, and risk differs by product and dose. The magnitude must be compared with age-specific disease risk rather than flattened into a universal slogan.

20. Breakthrough Infections Do Not Mean Zero Effectiveness

“Vaccinated people still get infected” is often presented as a decisive falsification. That would only follow if the claim being tested were “the vaccine prevents 100% of infections forever.” Few vaccines meet that standard, and many were never expected to. Effectiveness is a relative reduction in an outcome under specified conditions.

If an unvaccinated group has 100 hospitalizations per 100,000 people and a vaccinated group has 30, the vaccine reduced hospitalization risk substantially even though 30 vaccinated people were hospitalized. The existence of breakthroughs says nothing by itself about the size of the relative or absolute risk reduction.

Base rates can also create counterintuitive raw counts. If 95% of an elderly population is vaccinated, vaccinated people can constitute many or even most hospitalized patients while still having a much lower individual risk. Comparing numerators without denominators commits the base-rate fallacy. One must compare rates within comparable age and risk groups.

Protection can also differ by endpoint. Respiratory vaccines may be better at preventing severe lower-respiratory disease than every transient upper-airway infection, especially as mucosal antibodies wane and variants change. This distinction became highly visible with SARS-CoV-2 but is not unique to COVID-19.

Verdict: Breakthrough disease disproves perfect sterilizing immunity, not vaccine effectiveness. Rates, denominators, endpoints, and time since vaccination are required.

21. “The Flu Shot Barely Works”

Influenza vaccination is an excellent case where overpromising is unnecessary. Flu vaccines are imperfect. Influenza evolves quickly, strain selection occurs months before a season, manufacturing takes time, immune responses vary by age, and effectiveness can differ substantially from year to year and by subtype. Some seasons are genuinely disappointing.

Yet “imperfect” and “useless” are not synonyms. CDC estimates that in the high-severity 2024–2025 U.S. influenza season, vaccination prevented roughly 10 million illnesses, 5 million medical visits, 180,000 hospitalizations, and 12,000 deaths.^[26] These are modeled estimates rather than directly counted alternate-universe outcomes, so their precision should not be exaggerated. But the models incorporate measured vaccine coverage and effectiveness, and the scale demonstrates how moderate individual effectiveness can translate into large population benefit.

The flu shot also illustrates why “I got vaccinated and still got flu” is weak evidence. First, many illnesses colloquially called “the flu” are caused by other respiratory viruses. Second, no vaccine is perfectly effective. Third, a vaccinated person who truly develops influenza may still have reduced risk of severe outcomes even if infection was not prevented.

A serious critique of influenza vaccination can discuss mismatch, waning immunity, age-dependent effectiveness, manufacturing platforms, and the uncertainty of annual effectiveness estimates. None of those requires the false claim that flu vaccination has never prevented illness or hospitalization.

Verdict: Influenza-vaccine effectiveness is variable and sometimes modest, but substantial reductions in illness and severe outcomes are measurable across many seasons.

22. “Why Vaccinate a Newborn Against Hepatitis B?”

This objection often assumes hepatitis B is exclusively an adult sexually transmitted infection. HBV can indeed spread through sexual contact and blood exposure, but it can also pass from mother to infant around birth and through household blood exposure. The age of infection matters enormously because infants who become infected are much more likely to develop chronic infection than adults.

WHO reports that, without preventive intervention, mother-to-child transmission risk can reach 70%–90% when the mother has a high viral load or is HBeAg-positive. A completed hepatitis B vaccine series provides roughly 98%–100% protection, and WHO recommends a first dose as soon as possible after birth, preferably within 24 hours, as part of preventing perinatal transmission.^[27]

Why give a birth dose if prenatal screening exists? Screening greatly improves targeting but is not infallible: records can be missing, infection can be acquired after testing, lab results can be wrong, and household transmission can occur. A universal birth-dose strategy is therefore partly a systems-engineering decision intended to create a safety net. Whether a country should use universal versus risk-based birth dosing is a policy question that can change with epidemiology, screening reliability, and health-system design. The biology behind early protection is not mysterious.

Verdict: The newborn dose is not based on an assumption that infants are sexually active. It targets perinatal and early-childhood transmission, which carries a high risk of chronic HBV infection.

23. HPV Vaccination and Cancer Prevention

Human papillomavirus is a sexually transmitted virus, but the rationale for vaccination before sexual exposure is precisely that prophylactic vaccines work best before infection. Persistent infection with high-risk HPV types causes nearly all cervical cancers and contributes to anal, penile, vulvar, vaginal, and oropharyngeal cancers. Waiting until after exposure can reduce the preventable fraction.

Anti-HPV claims have included infertility, autoimmune disease, neurologic syndromes, and the moral argument that vaccination encourages sexual activity. The last is not a biomedical safety claim and must be evaluated separately. For reproductive safety, post-licensure monitoring has not established primary ovarian insufficiency as a consequence of HPV vaccination.^[21]

HPV vaccination also creates a powerful long-latency test of vaccine benefit. Precancerous lesions decline before invasive cancer because cancer takes years to develop. As vaccinated cohorts age, multiple countries have reported large reductions in cervical precancer and cervical cancer among women vaccinated at younger ages. That pattern is the opposite of what one would expect if the vaccine were medically useless.

Verdict: The fertility claim is unsupported, and the biological and population evidence supports HPV vaccination as a cancer-prevention intervention rather than merely an “STD shot.”

24. Vaccine Shedding and Vaccine-Derived Polio: The Real Exception

“Vaccine shedding is impossible” would be false. Some live attenuated vaccines contain organisms capable of limited replication. The oral polio vaccine is the clearest example. The attenuated poliovirus replicates in the intestine and can be excreted in stool. In settings with poor sanitation, that vaccine virus can spread to other people. Historically, this secondary spread could even help immunize contacts.

There is a real downside. In communities with low immunization coverage, vaccine-derived poliovirus can circulate for long enough to accumulate mutations that restore neurovirulence, producing circulating vaccine-derived poliovirus (cVDPV). WHO describes this mechanism explicitly.^[28] The older oral vaccine could also very rarely cause vaccine-associated paralytic poliomyelitis. In the United States, an average of about eight VAPP cases per year occurred during 1980–1999, around one per 2.4 million OPV doses distributed, before the country moved to an all-inactivated schedule.^[4]

This is an important correction to sloppy pro-vaccine rhetoric. Vaccine shedding is a real phenomenon for certain live vaccines. It does not follow that a non-live mRNA vaccine sheds replicating virus. Different platforms must remain different. The existence of VAPP also does not erase the enormous polio burden prevented by OPV; rather, it shows why eradication strategy eventually shifts toward vaccine formulations that cannot revert once wild-virus circulation is controlled.

Verdict: Shedding is real for some live attenuated vaccines. It is not a generic property of all vaccines, and it does not support claims of infectious shedding from non-live mRNA vaccines.

25. Real Vaccine Injuries and Historical Safety Failures

A credible analysis of anti-vaccine claims has to say something that simplistic reassurance often avoids: vaccines have caused serious injuries. Some products have been withdrawn because post-market evidence revealed risks that were not apparent before widespread use. Manufacturing failures have harmed people. These facts are not anti-vaccine; they are part of vaccine history.

The Cutter incident

In 1955, some lots of inactivated polio vaccine manufactured by Cutter Laboratories contained live poliovirus. More than 250 polio cases were attributed to the incident, including paralysis. The product was recalled, manufacturing and regulatory controls were tightened, and vaccination later resumed.^[29] No theory of vaccine safety should require pretending this did not happen.

1976 swine flu vaccine and Guillain-Barré syndrome

The 1976 U.S. swine-flu campaign was associated with roughly one additional case of GBS per 100,000 vaccine recipients, enough to contribute to suspension of the program while the association was investigated.^[29] Modern seasonal flu vaccines have shown either no increased risk in some seasons or much smaller elevations—typically around one to two additional GBS cases per million doses when a signal is present.^[19]

RotaShield and intussusception

The first U.S. rotavirus vaccine, RotaShield, was withdrawn in 1999 after post-licensure studies showed that it caused intussusception in some infants.^[29] Newer rotavirus vaccines were developed and are also associated with a small intussusception risk, estimated by CDC at approximately one additional case per 20,000 to 100,000 vaccinated U.S. infants.^[30]

MMR and febrile seizures

MMR is associated with a small increased risk of febrile seizure in young children, roughly one event per 3,000 to 4,000 vaccinated children in the relevant post-vaccination window.^[31] Febrile seizures are frightening but generally do not imply epilepsy or permanent neurological injury. The risk is still real and belongs in informed discussion.

Anaphylaxis

Severe allergic reactions can occur after vaccination. For many vaccines the rate is on the order of one per million doses, though it varies by product and population. This is one reason people are observed briefly after some vaccinations and why treatment for anaphylaxis is kept available.

These examples demolish two bad arguments simultaneously. They disprove “vaccines never cause serious adverse events.” They also disprove “regulators would never admit a vaccine problem.” Cutter led to changed manufacturing oversight. RotaShield was removed. GBS signals changed policy. Myocarditis changed COVID-vaccine recommendations in multiple places. The system is imperfect, but real harms have been detected precisely because safety surveillance exists.

Verdict: Vaccine injuries are real. The existence of documented product-specific harms is a reason for precise pharmacovigilance, not evidence that every alleged injury is causal.

26. “Unvaccinated Children Are Healthier”

This claim often rests on small surveys comparing families who choose vaccination with families who refuse it. Such groups differ in far more than vaccination. They may differ in health-care use, diagnosis seeking, socioeconomic status, breastfeeding, schooling, diet, geography, family size, religious community, smoking exposure, and willingness to consult physicians. If one group visits doctors less often, it can appear to have fewer diagnosed conditions even when underlying health is identical.

This is called confounding and ascertainment bias. It does not mean observational comparisons are worthless. It means a credible study must define the population in advance, obtain reliable vaccination records, adjust for major confounders, avoid self-selected internet samples, use objective outcomes where possible, and consider whether unvaccinated children are underdiagnosed because they interact with the medical system differently.

The strongest vaccine-effectiveness evidence also directly contradicts the idea that nonvaccination is globally healthier. Unvaccinated children are substantially more susceptible to measles, pertussis, varicella, and other targeted infections. Even if someone hypothesized a reduction in one unrelated chronic condition, that would have to be weighed against the demonstrated increase in vaccine-preventable disease rather than treated as an isolated benefit.

A legitimate research question is whether a particular schedule or vaccine is associated with a particular long-term outcome. “Unvaccinated children are healthier” is too vague to test until “healthier” is defined, the groups are made comparable, and diagnostic opportunity is addressed.

Verdict: Broad claims from self-selected vaccinated-versus-unvaccinated surveys are highly vulnerable to confounding and ascertainment bias. Disease-specific data show clear increases in vaccine-preventable infections among the unvaccinated.

27. “The Package Insert Lists Death, Autism, or Other Conditions, So the Vaccine Causes Them”

Package inserts are often treated online as secret confessions hidden in plain sight. Someone searches a manufacturer document, finds a frightening diagnosis in an adverse-events section, and concludes that the manufacturer has admitted causation. That is not how pharmacovigilance labeling works. A product label can contain several different kinds of information: adverse reactions demonstrated in controlled trials, events observed after approval, contraindications, warnings based on mechanism or class effects, and events reported voluntarily for which frequency and causality cannot be reliably determined.

This distinction is essential. If a stroke occurs after vaccination and is reported during post-market surveillance, the existence of that report can justify listing or monitoring without establishing that vaccination raised stroke risk. The same logic applies to drug labels. An event can be temporally associated with exposure without being caused by it. CDC's vaccine adverse-reaction guidance explicitly notes that adverse events following vaccination do not necessarily indicate causality and that more detailed epidemiological studies are often required to compare incidence among vaccinated and unvaccinated people.^[32]

A label is therefore a starting point for questions, not a substitute for comparative data. If an insert names a condition, the next questions should be: Which section is it in? Was it more common in the vaccine group than the control group? Was the signal statistically significant? Was it biologically plausible? Did active surveillance reproduce it? Did regulators determine causality, or was it merely reported after use?

There is an inverse mistake too. Manufacturers and regulators should not hide behind labeling language when a causal association is supported. Once evidence establishes a risk—such as myocarditis after certain mRNA COVID-19 vaccine doses—the risk belongs in warnings and benefit-risk decisions. The point is that the label must be read according to what its categories mean.

Verdict: Finding a diagnosis in a package insert does not by itself prove the vaccine causes it. The evidentiary meaning depends on the section and the data behind the listing.

28. Fetal Cell Lines: What Is Actually in the Vial?

This issue contains both a factual question and an ethical question, and confusing them helps nobody. Some vaccines have historically been produced using laboratory cell lines ultimately derived from fetal tissue obtained decades ago. In the United States, examples have included rubella, varicella, hepatitis A, and one rabies vaccine. These are long-maintained cell lines; production does not require a new abortion for each batch or each dose.^[33]

That does not mean a vaccine vial contains intact fetal tissue, organs, or fetal cells. Viruses or vaccine antigens are produced in a cell substrate and then purified. Depending on the manufacturing process, minute residual fragments of cellular DNA or protein can remain, just as trace process-related material can remain after other biological manufacturing. The safety question concerns the amount, fragment size, biological activity, and purification controls—not an image of recognizable fetal tissue being injected.

FDA explicitly evaluates cell substrates and residual DNA as part of biologics safety because theoretical risks such as adventitious viruses or biologically active DNA are real manufacturing questions.^[34] This is another useful distinction: acknowledging a theoretical risk is not the same as demonstrating clinical harm. Manufacturing standards limit and characterize residual material, and decades of use have not shown the sweeping cancer or genetic-alteration effects alleged from these cell lines.

There can still be a sincere moral objection. Someone may believe that use of a cell line descended from an elective abortion decades ago is ethically unacceptable regardless of whether the final vial contains fetal cells. That is a moral or theological argument, not a toxicology result. It should be debated honestly on those terms rather than converted into the false factual claim that a vaccine is “made of aborted babies.”

Verdict: Some vaccines are manufactured using historic human fetal-derived cell lines, but the final vaccines do not contain intact fetal tissue or fetal cells. Ethical objections are distinct from claims of biological toxicity.

29. SV40-Contaminated Polio Vaccine and Cancer Claims

SV40 is another case where the historical premise is real. Between 1955 and 1963, an estimated 10%–30% of U.S. polio vaccine doses were contaminated with simian virus 40 because monkey kidney cells used in production carried the virus. Once contamination was recognized, testing requirements were introduced to prevent SV40 from appearing in new vaccine lots.^[35]

Because SV40 can produce tumors in some experimental systems, researchers reasonably asked whether exposed vaccine recipients later developed excess cancers. This is exactly the sort of question long-term epidemiology is designed to answer. CDC summarizes the human evidence as generally reassuring, with most studies failing to find a causal association between receipt of SV40-contaminated polio vaccine and cancer.^[35] Modern vaccines do not contain SV40.

The lesson should not be “contamination never matters.” It plainly can. The lesson is that a historical contamination event must be separated into at least three questions: Was contamination present? Yes. Could the contaminant be biologically concerning in principle? Yes. Did exposed humans subsequently show the predicted cancer burden? The evidence has not established that causal link. Conflating those questions turns a genuine manufacturing failure into a claim stronger than the epidemiology supports.

Verdict: Historical SV40 contamination of some polio vaccines was real. The claim that it has been shown to cause a major human cancer epidemic is not supported by the bulk of epidemiological evidence.

30. Antibody-Dependent Enhancement and Dengvaxia: A Real Warning

Antibody-dependent enhancement, or related forms of vaccine-associated enhanced disease, are not imaginary concepts invented by vaccine critics. Under some immunological circumstances, prior antibodies can worsen a later infection rather than protect against it. Vaccine developers actively evaluate this possibility because it has mattered historically.

Dengue provides the clearest modern caution. Dengue has four major serotypes, and a second infection with a different serotype can carry a greater risk of severe disease. The first licensed dengue vaccine, Dengvaxia, performed differently depending on whether a recipient had previously been infected. Follow-up evidence showed an increased risk of severe dengue in vaccinated people who were dengue-seronegative at baseline. WHO's Global Advisory Committee on Vaccine Safety therefore recommended that Dengvaxia not be administered to people without evidence of previous wild dengue infection.^[36]

This episode matters for two reasons. First, it proves that “a vaccine can never enhance disease” is false. Second, it demonstrates what evidence of enhancement looks like: a defined product, a defined immunological subgroup, a biologically coherent mechanism, and an increased clinical risk that emerges in follow-up. That is very different from labeling any post-vaccination infection “ADE.”

During COVID-19 vaccine development, enhancement was a legitimate theoretical concern because of earlier coronavirus and respiratory-virus research. The relevant question was whether vaccinated people who later encountered SARS-CoV-2 developed disproportionately severe disease. Clinical trials and subsequent population experience did not show the pattern expected from clinically important ADE. Breakthrough infection, by itself, is not enhancement; the disease has to be made more severe because of the prior immune response.

Verdict: Enhanced disease is a real vaccine-development hazard and Dengvaxia is an important example. That makes ADE a testable safety concern, not a universal explanation for any infection after vaccination.

31. “Vaccine Manufacturers Have No Liability”

In the United States, vaccine liability is unusual, but “manufacturers have zero liability and cannot be sued” is too simple. The National Childhood Vaccine Injury Act created the Vaccine Injury Compensation Program (VICP), a no-fault alternative to ordinary tort litigation for covered vaccines. Petitioners generally pursue compensation through the U.S. Court of Federal Claims first. The system was designed in part because waves of litigation threatened vaccine supply and price stability.^[37]

The program is funded through an excise tax on covered vaccines and can compensate people for qualifying vaccine-related injuries or deaths. Some conditions on the Vaccine Injury Table receive a presumption of causation when diagnostic and timing criteria are met; other claims require evidence of causation. Settlements do not necessarily constitute government admission that a vaccine caused the alleged injury.^[38]

Liability protections are substantial, but they are not absolute. HRSA notes circumstances in which civil action may still be possible and explains that petitioners who exhaust VICP procedures can in some situations pursue litigation outside the program.^[38] COVID-19 countermeasure claims have also operated under a different compensation framework, so they should not be casually conflated with the ordinary VICP.

There is a legitimate policy debate here. One can ask whether compensation is generous enough, whether legal barriers are too high, whether manufacturers have adequate incentives to improve safety, or whether traditional tort exposure would destabilize supply. But none of those questions establishes whether MMR causes autism or aluminum causes chronic disease. Liability structure is evidence about law and incentives, not direct evidence of biomedical causation.

Verdict: U.S. vaccine manufacturers receive significant statutory liability protections for covered vaccines, but “complete immunity from all lawsuits” is inaccurate. The compensation system is a policy fact, not proof that every alleged vaccine injury is causal.

32. Pharmaceutical Profit, Regulators, and Conflicts of Interest

Pharmaceutical companies make money from vaccines. Regulators can make mistakes. Researchers can have conflicts of interest. Public-health institutions can communicate badly, become defensive, or preserve a policy longer than evidence warrants. None of those propositions should be controversial.

But “Pfizer profits from a vaccine” and “therefore the vaccine causes infertility” are logically different claims. Financial incentive raises the prior probability of biased behavior and justifies transparency, independent replication, data access, and adversarial review. It does not tell us whether a biological effect exists. Oil companies have conflicts regarding climate science; that does not mean every statement an oil-company scientist makes about atmospheric chemistry is automatically false. Evidence still has to be examined.

The strongest protection against institutional bias is not blind trust but distributed evidence. Vaccine questions are studied by universities, national health systems, independent researchers, competing companies, regulators in different countries, and databases covering millions of people. They do not always agree. When an adverse effect appears across independent systems—as happened with myocarditis—the signal becomes harder to dismiss. When a proposed effect repeatedly fails to appear across different systems—as with MMR and autism—that also matters.

Conflicts should therefore change how evidence is weighted, not replace evidence. A claim should not become true merely because the accused institution is capable of wrongdoing.

Verdict: Conflicts of interest are real reasons for scrutiny and replication, but they are not evidence that a particular medical claim is true.

33. Mandates Are a Policy Question, Not a Biological Result

People frequently collapse at least four questions into one: Does a vaccine reduce a disease outcome? What adverse effects does it cause? Should an individual receive it? Should the government, employer, school, military, or hospital require it? The fourth question does not automatically follow from the first three.

A person can conclude that a vaccine is effective and still oppose a mandate on civil-liberties grounds. Another can support a mandate in a narrow setting because the disease threatens vulnerable people. Reasonable policy analysis can consider transmission reduction, severity, duration of protection, age-specific risk, alternatives such as testing, natural immunity, administrative feasibility, and the coercive cost of enforcement.

During COVID-19, this distinction was often lost. Arguments about changing transmission effectiveness were treated as though they settled every ethical question, while moral objections were sometimes presented as proof that the vaccine itself was ineffective. Both are category errors.

Likewise, liability systems and compensation programs deserve separate examination. The existence of a compensation program does not prove vaccines are uniquely dangerous; neither does it prove every claimant was uninjured. These are legal mechanisms for allocating the cost of rare harms in a population-level program.

Verdict: Biomedical efficacy and safety data inform mandate policy but do not mechanically determine it. One can criticize coercive policy without inventing medical claims.

34. Claim-by-Claim Verdicts

Conclusion

The strongest case for vaccination does not require pretending vaccines are perfect. In fact, pretending that creates an avoidable weakness. The historical record contains manufacturing failures. Some vaccines have been withdrawn. Specific products can cause rare but serious adverse effects. Effectiveness can wane. Influenza vaccines can mismatch circulating strains. Acellular pertussis immunity is imperfect. Oral polio vaccine can, under particular conditions, generate vaccine-derived poliovirus. mRNA COVID-19 vaccination carries a real myocarditis risk concentrated in younger males. These are not concessions that destroy vaccine science. They are vaccine science.

What fails is the attempt to turn those specific facts into a universal theory in which vaccination is responsible for autism, infertility, SIDS, immune collapse, chronic disease, genetic alteration, and every medical event reported after an injection. Those claims make predictions. Autism should rise and fall with the alleged causal exposure. Infertility should appear in reproductive outcomes. SIDS should occur at excess rates after vaccination. Vaccinated cohorts should show the predicted chronic-disease burden. VAERS signals should survive controlled comparisons. Most of these predictions have been tested, often repeatedly, and the sweeping claims do not survive.

The reverse mistake is equally important. “Vaccines have saved millions of lives” does not imply that every vaccine, every dose, every age group, every mandate, or every recommendation is automatically correct forever. Benefit-risk analysis can change when pathogens evolve, when population immunity changes, when a safer product becomes available, or when new adverse-event evidence appears. Scientific confidence should be proportional to evidence, not institutional prestige.

The broader historical record remains difficult to evade. Measles incidence collapsed after measles vaccination. Polio paralysis collapsed after polio vaccination. Chickenpox hospitalizations and deaths collapsed in a highly sanitized modern country after varicella vaccination. Smallpox disappeared from natural circulation after an eradication campaign centered on vaccination and surveillance. These are not laboratory abstractions. They are population-scale changes with timelines, mechanisms, and independent replication.

So the correct position is not “vaccines are safe because experts say so.” It is more demanding: specific vaccine claims should be tested against specific evidence, using the same causal standards whether the result is reassuring or uncomfortable. Under that standard, many of the most common anti-vaccine claims fail. A smaller number identify real issues but exaggerate them beyond recognition. And a few point to genuine adverse effects that deserve to be stated plainly, quantified accurately, and incorporated into better medical decisions.

That is a stronger defense of evidence than either reflexive trust or reflexive distrust. It allows a vaccine to work without being perfect, an injury to be real without proving a conspiracy, a regulator to be fallible without making every database fraudulent, and a person to debate policy without rewriting biology.

References

1. CDC — Questions About Measles: U.S. measles burden before vaccination.
2. CDC — Measles Vaccine Recommendations: MMR effectiveness.
3. CDC — Pertussis Surveillance and Trends.
4. CDC — Manual for Surveillance of Vaccine-Preventable Diseases: Poliomyelitis.
5. Marin et al. — Decline in Severe Varicella Disease During the U.S. Vaccination Program.
6. WHO — Global immunization efforts have saved at least 154 million lives over 50 years.
7. Hviid et al. (2019) — MMR Vaccination and Autism: Nationwide Cohort Study.
8. WHO (2026) — Vaccines, thiomersal and autism spectrum disorder: evidence review 2010–2025.
9. CDC — Thimerosal and Vaccines.
10. FDA — Common Ingredients in FDA-Approved Vaccines.
11. CDC — Adjuvants and Vaccines.
12. Andersson et al. (2025) — Aluminum-Adsorbed Vaccines and Chronic Diseases in Childhood.
13. WHO GACVS — Review of vaccines, aluminum/thiomersal, and autism.
14. Iqbal et al. (2013) — Early Childhood Vaccine Antigen Exposure and Neuropsychological Outcomes.
15. National Academies (2013) — The Childhood Immunization Schedule and Safety.
16. FDA — Vaccine Development 101.
17. CDC — Vaccine Adverse Event Reporting System (VAERS).
18. CDC — Sudden Infant Death Syndrome and Vaccines.
19. CDC — Guillain-Barré Syndrome and Vaccines.
20. Zaçe et al. (2022) — COVID-19 Vaccines and Fertility: Systematic Review and Meta-analysis.
21. CDC — HPV Vaccine Safety and Reproductive Concerns.
22. CDC — COVID-19 Vaccine Basics: mRNA and DNA.
23. CDC Archive — COVID-19 Vaccine Myths and Facts: Shedding.
24. Patone et al. (2022) — Myocarditis After COVID-19 Vaccination and SARS-CoV-2 Infection by Age and Sex.
25. Karlstad et al. (2022) — SARS-CoV-2 Vaccination and Myocarditis in a Nordic Cohort of 23 Million.
26. CDC — Benefits of Flu Vaccination; 2024–2025 burden prevented.
27. WHO — Preventing Mother-to-Child Transmission of Hepatitis B.
28. WHO — Poliomyelitis: oral polio vaccine and circulating vaccine-derived poliovirus.
29. CDC — Historical Vaccine Concerns: Cutter, 1976 GBS, RotaShield, and other signals.
30. CDC — Rotavirus Vaccination and Intussusception Risk.
31. CDC — MMR Vaccine Safety and Febrile Seizures.
32. CDC — Preventing and Managing Adverse Reactions: temporal association versus causality.
33. Children's Hospital of Philadelphia — Vaccine Ingredients: Fetal Cells.
34. FDA — Safety Evaluation of Cell Substrates and Residual DNA in Vaccines.
35. CDC — Historical Vaccine Concerns: SV40 contamination.
36. WHO GACVS — Dengvaxia safety update and seronegative risk.
37. HRSA — About the National Vaccine Injury Compensation Program.
38. HRSA — VICP Frequently Asked Questions: liability protections, causation, and settlements.