Review of the 2002 Danish MMR Study (Madsen et al., 2002)

A Population-Based Study of Measles, Mumps, and Rubella Vaccination and Autism

Madsen et al

Published November 7, 2002

N Engl J Med 2002;347:1477-1482

DOI: 10.1056/NEJMoa021134

Review of the 2002 Danish MMR Study

The 2002 Danish cohort study by Madsen and colleagues is often cited as a cornerstone in the argument against any link between MMR vaccination and autism. The study followed over 537,000 children born between 1991 and 1998, making it the largest of its kind at the time. Its headline conclusion was clear: “This study provides strong evidence against the hypothesis that MMR vaccination causes autism” (Madsen et al., 2002, NEJM).

However, one of the most important criticisms centres on the fact that the study used diagnosis dates rather than the onset of symptoms. Autism is rarely diagnosed immediately. In fact, Madsen’s own data show the average age at diagnosis was “4 years 3 months for autistic disorder and 5 years 3 months for other autism spectrum disorders”. This creates a gap of nearly three years between when the vaccine is given (12–15 months) and when autism is officially recorded.

This lag of nearly three years is not incidental — it coincides exactly with how long it typically takes for children to move from first symptoms to a formal diagnosis. Parents often report noticing traits such as language delay or social withdrawal within months of vaccination, but because the study relied on diagnosis dates rather than onset of symptoms, such signals would have been lost in the data.

By relying on diagnosis rather than onset, the study design could not capture short-term changes or clustering of symptoms following vaccination. Many parents report noticing regression or first symptoms within months of the jab, yet this “signal” would be invisible when measured against much later diagnosis dates. As the authors themselves admitted, “We did not find any association between the age at the time of vaccination, the time since vaccination, or the date of vaccination and the development of autistic disorder”. But this negative finding may simply reflect the bluntness of the tool they chose, not the absence of an effect.

In addition, the study was conducted during a period when diagnostic criteria for autism were changing and broadening, which complicates interpretation. The authors noted that “the increase in autism diagnoses may be due to increased awareness, broadening of the diagnostic concept, or changes in referral patterns”. This makes it difficult to separate genuine incidence from reporting artefacts.

Conclusion

While the Danish study is impressive for its size and scope, its reliance on late diagnosis data makes its conclusions less solid than they first appear. The gap between vaccination (at 1 year) and diagnosis (at 4–5 years) means that any short-term associations would almost certainly be missed. Quoting the authors’ confident conclusion that “there is no association between MMR vaccination and autism” overlooks the fact that their methods may not have been sensitive enough to detect one.