Does the MMR vaccine cause autism? What the large studies actually found

The measles, mumps and rubella (MMR) vaccine does not cause autism. The hypothesised link has been examined in studies following hundreds of thousands of children, and it does not appear [s1]. The question keeps returning because the fear is genuine and the timing is suggestive — the MMR vaccine is given in early childhood, around the age when autism is typically first recognised — but coincidence in timing is exactly what large epidemiological studies are built to separate from cause.

A nationwide cohort of 657,461 children

The most comprehensive test is a Danish nationwide cohort study of 657,461 children born between 1999 and the end of 2010, followed from the age of one through August 2013 [s1]. Using national registries, the researchers linked each child's MMR-vaccination status to later autism diagnoses while adjusting for age, sex, birth year, other childhood vaccines, a sibling history of autism, and autism risk factors [s1].

Over 5,025,754 person-years of follow-up, 6,517 children were diagnosed with autism, an incidence rate of 129.7 per 100,000 person-years [s1]. Comparing vaccinated with unvaccinated children yielded a fully adjusted hazard ratio of 0.93 (95% CI, 0.85 to 1.02) [s1]. A hazard ratio of 1.00 would mean no difference; 0.93 sits just below it, with a confidence interval that comfortably includes 1.00 — no increased risk, and if anything a point estimate slightly under parity, which is best read as no effect rather than protection.

Crucially, the study went looking specifically for a subgroup that might be susceptible. It found no increased risk for autism after MMR vaccination in children with a sibling history of autism, in children with other autism risk factors, or clustered in any particular time period after vaccination [s1]. That matters because the "some vulnerable children" argument is where the hypothesis usually retreats once the overall numbers come back null.

An earlier cohort found the same thing

This was not a lone result. An earlier population-based study, also in Denmark, followed 537,303 children, of whom 440,655 — 82.0% — had received the MMR vaccine [s2]. Among them, 316 were diagnosed with autistic disorder and 422 with other autistic-spectrum disorders [s2]. After adjustment for confounders, the relative risk of autistic disorder in vaccinated versus unvaccinated children was 0.92 (95% CI, 0.68 to 1.24), and for other autistic-spectrum disorders 0.83 (95% CI, 0.65 to 1.07) [s2]. There was no association between the age at vaccination, the time since vaccination, or the calendar date of vaccination and the development of autistic disorder [s2]. Two large cohorts, run years apart in the same registry system, returned the same answer.

Pooling the evidence

A meta-analysis brought the studies together, combining five cohort studies totalling 1,256,407 children with five case-control studies of 9,920 children [s3]. In the cohort data, there was no relationship between autism and any vaccination (odds ratio 0.99, 95% CI 0.92 to 1.06), nor with the MMR vaccine specifically (0.84, 95% CI 0.70 to 1.01), nor with the preservative thimerosal (1.00, 95% CI 0.77 to 1.31) or its mercury component (1.00, 95% CI 0.93 to 1.07) [s3]. The case-control data agreed: grouped by exposure type, the odds ratio was 0.85 (95% CI 0.76 to 0.95) [s3]. Every estimate clusters around or below 1.00, and none points to harm.

The chart shows those cohort odds ratios with their confidence intervals. What a reader should take from it is not that vaccines prevent autism — the estimates hug the no-effect line and should be read as null — but that across more than a million children, the intervals around every exposure sit on the "no association" mark, including for the specific components, thimerosal and mercury, that separate versions of the theory have blamed.

What this does and does not settle

These are observational studies, and observational evidence cannot prove a negative with mathematical certainty; it can only fail, repeatedly and at enormous scale, to find an effect that a real cause would produce. That is what has happened here across independent populations, designs and analyses [s1][s2][s3]. None of this diagnoses or explains autism, whose causes are an active field of research, and none of it is guidance about an individual child's vaccinations — that belongs with a clinician who knows the child's history. What the evidence does close off is the specific claim that the MMR vaccine is a cause of autism.