Association Between Thimerosal-Containing Vaccine and Autism - 2003

Association Between Thimerosal-Containing Vaccine and Autism

Hviid et al

Published October 1, 2003

JAMA

2003;290;(13):1763-1766

doi:10.1001/jama.290.13.1763

In 2003, Hviid and colleagues published a large Danish cohort study in The Journal of the American Medical Association (JAMA), analysing 467,450 children to test whether thimerosal-containing vaccines increased the risk of autism. They concluded: “The risk of autism and other autism spectrum disorders did not differ significantly between children vaccinated with thimerosal-containing vaccines and those vaccinated with thimerosal-free vaccines.”

Critics, however, argue that the study was not sensitive enough to rule out a modest effect. The authors themselves admitted the estimate of risk carried uncertainty, reporting a rate ratio of 0.85 (95% CI, 0.60–1.20). This wide interval means children exposed to thimerosal could have been 40% less likely or up to 20% more likely to develop autism. In other words, the study could not exclude a possible increased risk — yet it was widely presented as definitive proof of no link.

Another concern is that, like earlier Danish studies, this research relied on diagnosis dates from health registries rather than the actual onset of symptoms. Since autism is often diagnosed years after first signs appear, the timing analysis may have blurred any temporal relationship between thimerosal exposure in infancy and the emergence of autistic traits.

Critics also point to how diagnostic criteria were changing rapidly in the 1990s. As the authors acknowledged: “The increase in autism diagnoses may be due to increased awareness, broadening of the diagnostic concept, or changes in referral patterns.” This raises the possibility that background shifts in diagnosis could have obscured subtle environmental contributions.

Conclusion

While the 2003 Danish study by Hviid et al. is often cited as strong evidence against a thimerosal–autism link, its findings are less definitive when viewed closely. The wide confidence intervals (0.60–1.20) mean the study cannot exclude the possibility of a modest increased risk. Its reliance on diagnosis dates, rather than the onset of symptoms, likely masked any short-term effects, particularly given the lag between early signs and formal diagnosis. Furthermore, rapid changes in autism awareness and diagnostic criteria during the study period complicate interpretation of the data. From a critics’ standpoint, the conclusion that thimerosal is not associated with autism is premature; instead, the study shows that the question remains unresolved, especially for rare or susceptible subgroups of children.