Vaccines do not cause autism. Here is every study, and the fraud that started the claim.

This is the third piece in our series on autism. In the first, we showed that the dramatic rise in autism diagnoses—from 1 in 150 children in 2000 to 1 in 31 in 2022—is mostly explained by changes in how autism is defined and how hard we look. In the second, we followed the money behind autism therapy. This piece takes on the claim at the centre of it all: that vaccines are the hidden cause.

To answer it properly, you need to understand what autism actually is, what vaccines actually do, how the claim started, and what happened when scientists tested it on a scale the original paper never imagined.

What autism is

Autism spectrum disorder (ASD) is a neurodevelopmental condition. It affects how a person communicates, interacts socially, and processes sensory information, and it often involves repetitive behaviours or intensely focused interests. It is a spectrum: some autistic people need round-the-clock support; others live independently and may not be diagnosed until adulthood.

Autism is not a disease you catch. It is not something that happens to a healthy brain and damages it. Brain-imaging and genetic studies consistently show that the differences are present before birth. A 2019 study using MRI scans of children later diagnosed with autism found structural brain differences detectable at six months of age—months before the MMR vaccine is given. Twin studies put the heritability of autism between 64% and 91%. Hundreds of genes contribute small amounts of risk, many of them active during prenatal brain development.

None of this means autism is purely genetic. Environmental factors during pregnancy—older parental age, very premature birth, certain prenatal infections, prenatal exposure to valproate—modestly increase risk. But the key word is prenatal. The brain architecture associated with autism is being laid down in the womb, not at the 12-month well-child visit.

As we showed in our first article, the reason autism appears to have exploded is mostly that the definition has expanded enormously and we got much better at finding it. In 2000 we counted 1 in 150 children. By 2022, 1 in 31. The children didn't change that much. The criteria did.

What vaccines are, and what they do

A vaccine teaches the immune system to recognise a specific pathogen—a virus or bacterium—without causing the disease. It does this by exposing the body to a harmless piece of the pathogen (a protein, a killed virus, or a weakened live virus) so that the immune system builds memory. If the real pathogen arrives later, the body already knows how to fight it.

The US childhood immunisation schedule recommends vaccines against 14 diseases by age two, given across roughly a dozen visits. The schedule has grown over the decades:

• 1983: children received vaccines against 7 diseases (diphtheria, tetanus, pertussis, polio, measles, mumps, rubella).
• 2000: the schedule covered 11 diseases, adding hepatitis B, Haemophilus influenzae type b (Hib), varicella (chickenpox), and hepatitis A.
• 2026: the schedule covers 14 diseases, having added pneumococcal disease, rotavirus, and influenza.

More vaccines sounds like more immune challenge. But the actual immune load has gone down. The old whole-cell pertussis vaccine alone contained about 3,000 bacterial proteins. The entire 2026 schedule contains roughly 160 immunological components (antigens) total—far fewer than children encountered in the 1980s. A child's immune system handles thousands of new antigens every day just from food, dirt, and the bacteria on their skin.

The paper that started it

In February 1998, Andrew Wakefield, a British gastroenterologist, published a paper in The Lancet with 12 co-authors. It described 12 children who had been referred to his clinic with gastrointestinal symptoms and developmental regression. The parents of 8 of the 12 reported that symptoms began shortly after the child received the MMR (measles-mumps-rubella) vaccine.

The paper did not actually prove a link. It was a case series—a description of 12 children, with no control group and no statistical analysis. But at a press conference on the day of publication, Wakefield went beyond the paper: he recommended that the combined MMR vaccine be replaced with three separate shots, a year apart.

The claim landed in a public already anxious about a vaccine schedule that had recently added several new shots. British media coverage was enormous. MMR uptake in England dropped from 92% in 1996 to 80% in 2003, and below 70% in parts of London. Measles, which had been nearly eliminated, came back. In 2006, a 13-year-old boy in England became the first person to die of measles there in 14 years.

How the fraud unravelled

Almost immediately, other researchers tried to replicate Wakefield's findings. None could. But it took investigative journalism to expose what had actually happened inside the study.

1. 1998The Lancet publishes Wakefield's paper on 12 children. At a press conference, Wakefield recommends splitting the MMR into single vaccines.
2. 2004Investigative journalist Brian Deer reveals in The Sunday Times that Wakefield had been paid £435,643 by a lawyer preparing a lawsuit against MMR manufacturers—before the study began. Wakefield had not disclosed this to The Lancet. Ten of his 12 co-authors publish a retraction of the paper's interpretation, stating: “We wish to make it clear that in this paper no causal link was established between MMR vaccine and autism.”
3. 2007Deer publishes further findings: the children in Wakefield's study had been recruited through anti-vaccine campaigners, not referred in the normal way. Their medical records, when compared to the paper, showed discrepancies in the timelines of when symptoms reportedly began.
4. 2010 · JanuaryThe UK General Medical Council finds Wakefield guilty of serious professional misconduct on more than 30 charges, including subjecting children to invasive medical procedures (colonoscopies, lumbar punctures) that were not clinically indicated, without proper ethical approval.
5. 2010 · FebruaryThe Lancet fully retracts the paper, stating that “several elements” of it “are incorrect.”
6. 2010 · MayWakefield is struck off the UK medical register and can no longer practise medicine in Britain.
7. 2011Deer publishes a series in the BMJ documenting that the study was not just flawed but deliberately fraudulent. He shows, child by child, that Wakefield altered the clinical histories of the 12 children: timelines were changed, pre-existing conditions were omitted, and normal biopsy results were reported as abnormal. The BMJ's editor calls it “an elaborate fraud.”

No other researcher has ever reproduced Wakefield's findings. The paper had 12 children and no controls. What came next had millions of children and rigorous designs.

What two million children showed

After 1998, researchers around the world ran large studies specifically to test the vaccine-autism hypothesis. Every one came back negative. Here are the largest and most important.

MMR and autism

• Denmark, 2002 (537,303 children). Madsen et al. followed every child born in Denmark from 1991 to 1998. Children who received the MMR vaccine were no more likely to develop autism than children who did not. The study was large enough to detect even a small increase in risk, and found none.
• United Kingdom, 2004 (5,000+ cases). Smeeth et al. used the UK's General Practice Research Database. Children with autism were no more likely to have received MMR than matched controls. There was no clustering of diagnoses after vaccination.
• Japan, 2005. Japan withdrew the MMR vaccine entirely in 1993 after concerns about the mumps component (unrelated to autism). If MMR caused autism, rates should have fallen. Honda, Shimizu, and Rutter studied children in Yokohama and found that autism diagnoses continued to rise after MMR was withdrawn—a natural experiment that directly contradicts the claim.
• United States, 2015 (95,727 children). Jain et al. published in JAMA the largest US study, which specifically included children who had an older sibling with autism—the group most often cited by vaccine-hesitant parents. MMR vaccination was not associated with increased autism risk, even among children already at higher genetic risk.
• Denmark, 2019 (657,461 children). Hviid et al. followed every child born in Denmark from 1999 to 2010 through 2013. The hazard ratio for autism among vaccinated children was 0.93—meaning MMR-vaccinated children were, if anything, slightly less likely to be diagnosed with autism, though the difference was not statistically meaningful. The study explicitly tested and rejected the idea that MMR triggers autism in susceptible subgroups.
• Meta-analysis, 2014 (1.27 million children). Taylor, Swerdfeger, and Eslick combined data from five cohort studies and five case-control studies across multiple countries. Their pooled result: no relationship between vaccines and autism, between MMR and autism, or between thimerosal and autism.

Thimerosal and autism

After the MMR claim failed to hold up, concern shifted to thimerosal, a mercury-based preservative used in some multi-dose vaccine vials. Thimerosal contains ethylmercury, which the body clears in days—unlike methylmercury (the kind in fish), which accumulates. But the names sound similar, so the fear transferred easily.

In 1999, as a precaution and to reduce overall mercury exposure, the CDC and AAP recommended removing thimerosal from childhood vaccines. By 2001, thimerosal had been removed or reduced to trace amounts in all routinely recommended childhood vaccines except multi-dose influenza vials (a thimerosal-free version is available).

If thimerosal caused autism, removing it should have caused autism rates to fall. They didn't.

• Denmark, 2003 (467,450 children). Hviid et al. compared children who received thimerosal-containing vaccines with those who received thimerosal-free versions. No difference in autism rates.
• United States, 2010. Price et al. used the Vaccine Safety Datalink and found no association between prenatal or early-life thimerosal exposure and autism.
• CDC prevalence data, 2000–2022. Thimerosal was removed from childhood vaccines between 1999 and 2001. Children born after 2001 had virtually no thimerosal exposure from vaccines. The autism rate among those children went from 1 in 150 to 1 in 31. If anything, it accelerated.

Too many vaccines too soon?

When individual vaccine ingredients failed as explanations, the claim shifted again: maybe it was the number of vaccines, or the schedule, overwhelming young immune systems.

• United States, 2013 (1,008 children). DeStefano et al. at the CDC compared the total number of antigens received in the first two years of life between children with and without autism. There was no association between cumulative antigen exposure and autism risk.
• Immunological reality. As noted above, the total antigen load from the entire modern vaccine schedule (~160 antigens) is a tiny fraction of what a child's immune system processes daily. A single upper respiratory infection exposes a child to 4 to 10 new antigens. The “too many too soon” hypothesis has no biological mechanism and no supporting data.

The timeline, side by side

One reason the vaccine claim feels intuitive is timing. The first dose of MMR is given at 12 to 15 months. Autism symptoms are typically first noticed by parents between 12 and 24 months. The two events land in the same window, and the human brain is wired to assume that when B follows A, A caused B.

But correlation in timing is not causation. Here is the full timeline of vaccination and autism diagnosis in the United States, placed side by side.

1. 1943Leo Kanner publishes the first clinical description of autism, years before widespread childhood vaccination programmes exist.
2. 1963The first measles vaccine is licensed in the US.
3. 1971The combined MMR vaccine is introduced.
4. 1980DSM-III creates “infantile autism” as its own diagnosis for the first time. The definition is narrow: severe, early onset, profound social withdrawal.
5. 1986Congress passes the National Childhood Vaccine Injury Act, creating a no-fault compensation system. The childhood schedule covers 7 diseases.
6. 1987DSM-III-R broadens the autism definition and adds PDD-NOS, a catch-all for children on the border.
7. 1991Autism becomes a category in US special-education law. Schools begin identifying and counting autistic children for the first time.
8. 1993Japan withdraws the MMR vaccine. Autism rates in Japan continue to rise.
9. 1994DSM-IV adds Asperger's disorder, covering people with autistic traits but no language delay. Autism is now a family of diagnoses. The vaccine schedule has expanded to 10 diseases.
10. 1998Wakefield publishes his paper on 12 children. UK MMR uptake begins to fall.
11. 1999The CDC and AAP recommend removing thimerosal from childhood vaccines as a precaution. Removal is completed by 2001.
12. 2000The CDC's ADDM Network begins counting: 1 in 150 eight-year-olds have autism. These children were born in 1992, before the vaccine schedule expanded much.
13. 2001Thimerosal is removed or reduced to trace amounts in all routine childhood vaccines except some flu shots.
14. 2002Denmark: 537,303 children, no MMR–autism link (Madsen et al.).
15. 2003Denmark: 467,450 children, no thimerosal–autism link (Hviid et al.).
16. 2004The Institute of Medicine (now the National Academies) reviews the full body of evidence and concludes: “The evidence favors rejection of a causal relationship between the MMR vaccine and autism” and between thimerosal-containing vaccines and autism. Ten of Wakefield's co-authors retract their interpretation.
17. 2005Japan: autism continued to rise after MMR was withdrawn (Honda et al.).
18. 2007The AAP recommends screening every child for autism at 18 and 24 months. More screening means more diagnoses.
19. 2010The Lancet fully retracts Wakefield's paper. Wakefield loses his UK medical licence.
20. 2013DSM-5 merges everything into one “autism spectrum disorder.” CDC: no link between number of vaccine antigens and autism (DeStefano et al.).
21. 2014Meta-analysis of 1.27 million children: no link (Taylor et al.).
22. 2015JAMA: 95,727 US children, including siblings of autistic children: no link (Jain et al.).
23. 2019Denmark: 657,461 children: no link (Hviid et al.).
24. 2022CDC ADDM Network: autism prevalence is now 1 in 31—in children who received vaccines with essentially no thimerosal, the same MMR formulation as two decades earlier, and a schedule that was studied in every combination tested.

The children born after thimerosal was removed in 2001 are now in their twenties. They were vaccinated, and they have more autism diagnoses than the generation before them—because, as we showed in our first article, the definition kept widening, screening got better, and more communities gained access to evaluations.

Why the myth persists

With this much evidence, why does the belief survive? Several reasons.

• The timing coincidence is powerful. Parents notice developmental differences around the same age vaccines are given. When you are watching your child regress, and the only new thing you can point to is a shot two weeks ago, the connection feels undeniable. That feeling is real and deserves empathy. But the science shows that the regression was going to happen regardless of the vaccine.
• The original paper got more coverage than its retraction. Wakefield's press conference was front-page news for weeks. The retraction, 12 years later, was a one-day story. Many people heard the claim and never heard the correction.
• Distrust of institutions is high. When people distrust pharmaceutical companies, regulators, or the medical system—and there are sometimes good reasons to—a claim that “they” are hiding something becomes more believable than the official position. But this is not a question of trusting one authority. It is dozens of independent research groups in seven countries, using different methods, arriving at the same answer.
• Autism still doesn't have a simple cause. Hundreds of genes, several prenatal environmental factors, and no single clear trigger. That uncertainty leaves a space that a clean, single-cause story fills. “It was the vaccine” is easier to hold than “it's a complex interaction of genetics and prenatal environment that we're still mapping.”
• Wakefield is still campaigning. Despite losing his licence and having his study branded as fraud, Wakefield has built a career as an anti-vaccine campaigner. He directed an anti-vaccine film in 2016 and remains active on social media. The claim has become an identity, not a hypothesis.

What it has cost

The vaccine-autism myth has not been harmless.

• In England, MMR uptake fell from 92% to 80% after the scare. Measles came back. Children were hospitalised. A boy died.
• In the US, measles was declared eliminated in 2000. Outbreaks returned: 667 cases in 2014, 1,282 in 2019—the highest in 27 years—concentrated in communities with low vaccination rates.
• Globally, the WHO listed “vaccine hesitancy” as one of the top ten threats to global health in 2019.
• Research funding that could have gone toward understanding what actually causes autism was diverted for over a decade to re-test a hypothesis that had already been debunked.
• Parents of autistic children have been sold unproven and sometimes dangerous “treatments” based on the false premise that their child was vaccine-injured—including chelation therapy, which removes heavy metals from the body and can cause kidney failure.

What actually does contribute to autism

Science hasn't found one cause because there isn't one. Autism arises from a combination of factors, nearly all of them operating before birth.

• Genetics: twin studies show heritability of 64–91%. Hundreds of genes, each contributing small amounts of risk, have been identified. Some cases involve rare, high-impact mutations.
• Older parental age: both older fathers and older mothers are associated with modestly higher risk, likely through increased rates of spontaneous genetic mutations.
• Prenatal factors: very premature birth, low birth weight, maternal infections during pregnancy, and prenatal exposure to the drug valproate all increase risk modestly.
• Brain development timing: the differences associated with autism are present in brain structure before birth and visible on MRI by six months of age—before most vaccines are given.

What is not on the list, after more than two million children studied: vaccines, thimerosal, the number of vaccines, or the vaccine schedule.

The bottom line

The claim that vaccines cause autism began with a fraudulent 1998 paper on 12 children, written by a researcher who was secretly paid by lawyers suing vaccine manufacturers, who subjected children to unnecessary invasive procedures without ethical approval, and who altered their medical records to fit his hypothesis. His paper was retracted. He lost his licence. His co-authors disowned his interpretation.

In the years since, studies covering more than two million children in Denmark, the United Kingdom, the United States, Japan, and elsewhere have tested the claim from every angle—MMR, thimerosal, total antigens, the schedule itself. Not one has found a link.

Meanwhile, the rise in autism diagnoses is well explained by changes in how we define and look for it. And autism itself, we now know, begins developing in the brain before a child is born—long before any vaccine is given.

Vaccines prevent measles, meningitis, whooping cough, and a dozen other diseases that used to kill and disable children routinely. They do not cause autism. That is not an opinion. It is what more than two decades of evidence, across millions of children, in countries on four continents, consistently shows.

If you have questions about your child's vaccine schedule, talk to your paediatrician. If you're concerned about your child's development, ask for a developmental screen. Both of those conversations matter. Neither one requires choosing between them.

Sources

1. Wakefield AJ, et al. Ileal-lymphoid-nodular hyperplasia, non-specific colitis, and pervasive developmental disorder in children. Lancet. 1998;351(9103):637–641. RETRACTED.
2. Murch SH, et al. Retraction of an interpretation. Lancet. 2004;363(9411):750.
3. The Editors of The Lancet. Retraction—Ileal-lymphoid-nodular hyperplasia, non-specific colitis, and pervasive developmental disorder in children. Lancet. 2010;375(9713):445.
4. Deer B. How the case against the MMR vaccine was fixed. BMJ. 2011;342:c5347.
5. Madsen KM, et al. A population-based study of measles, mumps, and rubella vaccination and autism. N Engl J Med. 2002;347(19):1477–1482.
6. Hviid A, Stellfeld M, Wohlfahrt J, Melbye M. Association between thimerosal-containing vaccine and autism. JAMA. 2003;290(13):1763–1766.
7. Smeeth L, et al. MMR vaccination and pervasive developmental disorders: a case-control study. Lancet. 2004;364(9438):963–969.
8. Honda H, Shimizu Y, Rutter M. No effect of MMR withdrawal on the incidence of autism: a total population study. J Child Psychol Psychiatry. 2005;46(6):572–579.
9. Price CS, et al. Prenatal and infant exposure to thimerosal from vaccines and immunoglobulins and risk of autism. Pediatrics. 2010;126(4):656–664.
10. DeStefano F, Price CS, Weintraub ES. Increasing exposure to antibody-stimulating proteins and polysaccharides in vaccines is not associated with risk of autism. J Pediatr. 2013;163(2):561–567.
11. Taylor LE, Swerdfeger AL, Eslick GD. Vaccines are not associated with autism: an evidence-based meta-analysis of case-control and cohort studies. Vaccine. 2014;32(29):3623–3629.
12. Jain A, et al. Autism occurrence by MMR vaccine status among US children with older siblings with and without autism. JAMA. 2015;313(15):1534–1540.
13. Hviid A, Hansen JV, Frisch M, Melbye M. Measles, mumps, rubella vaccination and autism: a nationwide cohort study. Ann Intern Med. 2019;170(8):513–520.
14. Institute of Medicine. Immunization Safety Review: Vaccines and Autism. Washington, DC: The National Academies Press; 2004.
15. CDC. Data and Statistics on Autism Spectrum Disorder (ADDM Network prevalence, 2000–2022).
16. Tick B, Bolton P, Happé F, Rutter M, Rijsdijk F. Heritability of autism spectrum disorders: a meta-analysis of twin studies. J Child Psychol Psychiatry. 2016;57(5):585–595.
17. Gaugler T, et al. Most genetic risk for autism resides with common variation. Nat Genet. 2014;46(8):881–885.
18. Hazlett HC, et al. Early brain development in infants at high risk for autism spectrum disorder. Nature. 2017;542(7641):348–351.
19. CDC. Recommended child and adolescent immunization schedule, United States, 2026.
20. Offit PA, Quarles J, Gerber MA, et al. Addressing parents' concerns: do multiple vaccines overwhelm or weaken the infant's immune system? Pediatrics. 2002;109(1):124–129.