Overview
The claim has two parts, and they are usually run together. The first is that vaccines cause autism. The second is that this was known and concealed.
The first has been tested about as thoroughly as any question in modern epidemiology, in cohorts running past a million children, and has not been supported. The second is more interesting, because the concealment turned out to run the other way: what was concealed was that the doctor who raised the alarm was being paid by lawyers preparing to sue vaccine manufacturers, and had filed a patent on a rival product.
It is worth separating three things that are routinely collapsed into one. A journal retracted a paper. A medical tribunal found dishonesty and struck a doctor off. And a body of subsequent research found the hypothesis unsupported. Those are three different findings by three different institutions, and they do not mean the same thing.
What the 1998 paper actually said
On 28 February 1998 The Lancet published a case series by thirteen authors, led by Andrew Wakefield, describing twelve children — eleven boys, mean age six — referred to a paediatric gastroenterology unit with a history of normal development followed by loss of acquired skills.
The paper reported intestinal abnormalities in all twelve. On the vaccine, its wording was careful: onset of behavioural symptoms was associated, by the parents, with MMR vaccination in eight of the twelve, with measles infection in one, and with an ear infection in another. Its stated conclusion was that the team had identified gastrointestinal disease and developmental regression "generally associated in time with possible environmental triggers."
That is a report of a temporal association described by parents. It is not a demonstration of cause, and the paper did not claim to be one. The causal claim was made publicly, around the paper, rather than in it — and it is the interpretation, not the paper, that ten of the thirteen authors retracted in March 2004.
What was not disclosed
In February 1996 — two years before publication — Wakefield had agreed to act as an expert in litigation against MMR manufacturers. The General Medical Council's panel found that he accepted £50,000 procured through the claimants' solicitor to pursue research, of which £25,000 was used for purposes other than those he had stated. He did not disclose this to the ethics committee or to the journal.
He also did not disclose to the journal that he was the inventor named on a patent for a rival measles vaccine, which the application claimed would additionally treat inflammatory bowel disease.
The panel quoted the journal's own disclosure test back at him: "Is there anything that would embarrass you if it were to emerge after publication and you had not declared it?"
The tribunal
The GMC's Fitness to Practise Panel delivered its findings on the facts on 28 January 2010 and its finding of serious professional misconduct, with a direction for erasure, on 24 May 2010. Those are two separate events four months apart, and accounts that compress them into a single January striking-off have it wrong.
The findings are specific. Wakefield held an honorary consultant post on the condition that he have no involvement in clinical management of patients; on five occasions he ordered investigations on children anyway, without paediatric qualifications, which the panel considered "a breach of trust of patients and employers alike."
On nine of eleven children he caused research to be undertaken without ethics committee approval, and acted contrary to each child's clinical interests. Three young children underwent lumbar puncture for research purposes when it was not clinically indicated.
On the paper's description of the children as "consecutively referred," the panel found his failure to disclose the true position "dishonest, in that his failure was intentional and that it was irresponsible." Given two later opportunities to correct the record — a published letter and a statement at a Medical Research Council meeting — he repeated it, and the panel found each repetition "dishonest and irresponsible."
There is also an episode at a birthday party, where he took blood from children in what the panel called "an inappropriate social setting," without ethics approval, paying them £5 each. He later described it in humorous terms in a public talk and said he intended to do it again. The panel found he had shown "callous disregard for any distress or pain the children might suffer."
One word the panel did not use is fraud. It found dishonesty, irresponsibility and serious professional misconduct. The stronger word entered later, through the BMJ.
The retraction, and how narrow it is
The Lancet retracted the paper in full on 2 February 2010, five days after the GMC's findings on the facts. The notice is short enough to quote whole in substance: several elements of the paper are incorrect, and in particular the claims that children were "consecutively referred" and that investigations were "approved" by the ethics committee "have been proven to be false."
That is the entire basis. The journal retracted on two specific falsehoods about how the study was conducted. It did not find that the clinical or pathological data had been fabricated, and a retraction is not a finding that the opposite of the paper is true.
The correction that is usually left out
John Walker-Smith was the senior clinician among the authors and was erased from the register alongside Wakefield in 2010. In March 2012 he appealed to the High Court and won; the findings against him were quashed.
This matters for the cover-up framing in both directions. It is evidence that the disciplinary process was not a foregone conclusion, since one of the two doctors it erased successfully overturned it. It is also the single most commonly omitted fact in accounts written from either side.
The other strand: thimerosal
Running alongside the MMR story is a separate claim about thimerosal, a mercury-containing preservative used in some vaccines.
On 1 July 1999 the US Food and Drug Administration asked manufacturers for plans to remove it. A week later the Public Health Service and the American Academy of Pediatrics issued a joint statement which is worth reading for what it actually says: the removal was precautionary, taken "because of scientific uncertainty at the time," and the same statement continued to recommend that all children be immunised on schedule, on the grounds that "the risks of not vaccinating children far outweigh the unknown and much smaller risk, if any, of exposure to thimerosal-containing vaccines."
A precaution taken under uncertainty was widely read afterwards as an admission of harm. The FDA's current position is that the evidence collected since "does not show any evidence of harm, including serious neurodevelopmental disorders, from use of thimerosal in vaccines," and that all vaccines routinely recommended for children aged six and under are available in thimerosal-free formulations.
The removal also produced a natural experiment. If thimerosal were driving autism diagnoses, taking it out should have brought them down. Denmark discontinued it in 1992; diagnoses continued to rise, including among children born after it was withdrawn. In California, where exclusion was accelerated between 1999 and 2001, researchers examining the state's developmental services caseload through March 2007 concluded that the data "do not show any recent decrease in autism in California despite the exclusion of more than trace levels of thimerosal from nearly all childhood vaccines."
What the large studies found
The epidemiology is unusually good, because Denmark's population registers allow a whole birth cohort to be followed.
A study of 537,303 Danish children born between 1991 and 1998, covering more than two million person-years, found an adjusted relative risk of autistic disorder among the vaccinated of 0.92, with a confidence interval spanning 0.68 to 1.24. Its authors described the result as "strong evidence against the hypothesis that MMR vaccination causes autism."
A second Danish cohort, published in 2019, followed 657,461 children born between 1999 and 2010 across more than five million person-years and recorded 6,517 autism diagnoses. The fully adjusted hazard ratio was 0.93, with a confidence interval from 0.85 to 1.02.
Both results sit slightly below 1, which is to say that vaccinated children in these cohorts were marginally less likely to be diagnosed — a difference too small to mean anything, and pointing in the opposite direction to the claim.
The vaccine court
Between 2007 and 2010 the United States Court of Federal Claims ran an Omnibus Autism Proceeding, established in 2002, to test the general causation theories in one place. Petitioners advanced three; the third was withdrawn. Each of the remaining two was tried through three test cases.
The scale is worth stating, because the proceeding is often described as a formality. The three cases in the first theory generated 5,000 pages of transcript, 939 medical articles — against about ten in a typical vaccine case — 50 expert reports and 28 testifying experts.
The decisions in the first theory were issued on 12 February 2009 and in the second on 12 March 2010. All six went against the petitioners. Special Master Hastings, deciding the lead case, wrote that the evidence was "overwhelmingly contrary to the petitioners' contentions."
His conclusion is also the fairest thing written about the families in the whole record, and it belongs here:
Nor do I doubt that Michelle's parents and relatives are sincere in their belief that the MMR vaccine played a role... I am convinced that the reports and advice given to the Cedillos by Dr. Krigsman and some other physicians... have been very wrong. Unfortunately, the Cedillos have been misled by physicians who are guilty, in my view, of gross medical misjudgment... I conclude that the Cedillos filed this Program claim in good faith.
The whistleblower episode
In 2014 the cover-up claim acquired a second act. A reanalysis of data from a 2004 CDC study reported an elevated risk among African American boys vaccinated before 36 months, and a CDC scientist, William Thompson, was said to have described the omission of that finding.
The reanalysis was published in August 2014, given an expression of concern three weeks later, and retracted on 3 October 2014. The retraction notice states its grounds precisely: "there were undeclared competing interests on the part of the author which compromised the peer review process," and post-publication review "raised concerns about the validity of the methods and statistical analysis."
The CDC's own statement, issued in August 2014, is more useful than either side's summary. It explains that the original study presented results both for all recruited children and for the subset with a Georgia birth certificate, because birth-certificate records carried fuller data on race and other characteristics; that the association it found — vaccination between 24 and 36 months being slightly more common among children with autism — was most likely explained by immunisation requirements for preschool special-education enrolment; and, notably, that "the data CDC collected for this study continue to be available for analysis by others."
What followed
Measles was declared eliminated from the United States in 2000, meaning no sustained domestic transmission. The CDC recorded 1,274 cases in 2019, 285 in 2024, 2,289 in 2025, and 3,134 by early September 2026.
That trajectory is the part of this story with a body count, and it is why the distinction the article opens with matters. A paper was retracted for two specific procedural falsehoods; a doctor was found dishonest for concealing who was paying him; and a hypothesis was tested at enormous scale and not supported. None of those three findings requires the others, and none of them is the same as saying the parents in that first case series were wrong about what they saw.