Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Thiomersal not linked to autism

Study
Results
Comment

Thiomersal (UK spelling) or thimerosal (US spelling) is an organic compound that contains ethylmercury, and which is frequently used as a preservative in chemistry and biochemistry. It has also been used as a preservative in vaccines.

A consequence has been a concern that giving organic mercury compounds to children might adversely affect their development, because high doses of mercuric compounds affect kidneys and nerves. Methylmercury has particularly been associated with environmental contamination and major human health problems. A study that examines the relationship between use of thiomersal and autism [1] is particularly welcome if it is sufficiently large and of high enough quality to answer the question with some authority.

Study

The study was conducted in Denmark, where, since 1968, people have had a unique identification number. This, together with the use of other specialist registries was used to construct a database of a comprehensive cohort of children born between 1990 and 1996. Information on vaccination could be linked to diagnosis of autism or autistic spectrum disorders, and potential confounders. Autism was diagnosed according to strict diagnostic criteria. All diagnoses up to the end of 2000 were recorded.

From 1970, the only thiomersal-containing vaccine used in Denmark was a whole-cell pertussis vaccine. This was used until March 1992 when the last batch was released and used. The vaccine was reformulated without thiomersal, and used until January 1997. The vaccine was administered at five weeks, nine weeks, and 10 months, irrespective of thiomersal content, and had the equivalent of 25 μg ethylmercury in the first dose and 50 μg in succeeding doses, for a maximum dose of 125 μg for each child.

This therefore constituted a population based cohort study of thiomersal use in childhood vaccination.

Results

In the cohort were 467,450 children with just under three million person-years of follow up. There were 440 cases of autism where the mean age at diagnosis was 4.7 years and 878 cases of other autistic spectrum disorders where the mean age at diagnosis was 6.0 years. Information was lost on 5,770 children (1.2%), mainly because of emigration.

Of the cohort, 95.6% were vaccinated at least once, 89% twice, and 63% received all three doses of the whole-cell pertussis vaccine. Only 4.4% did not receive any whole-cell pertussis vaccine.

There was no association between use of thiomersal and the risk of developing autism or autistic spectrum disorder (Table 1). For neither autism nor autistic spectrum disorders was there any increased diagnosis associated with use of thiomersal-containing vaccine. Neither was there any dose-response with increasing exposure to ethylmercury. The relative risks were adjusted for a range of possible confounders, but crude rates were no higher, and even lower, in children exposed to ethylmercury compared with those not exposed.


Table 1: Associations between use of thiomersal and autism or autistic spectrum disorders


Autism
Vaccination
Person-years at risk
Cases
Cases/10,000
Relative risk (95% CI)
All thiomersal-free
1,660,159
303
1.8
1.0
Any containing thiomersal
1,220,006
104
0.9
0.9 (0.6 to 1.2)
Dose of thiomersal
None
1,660,159
303
1.8
1.0
25 µg ethylmercury
169,920
18
1.1
1.0 (0.6 to 1.7)
75 µg ethylmercury
447,973
33
0.7
0.7 (0.5 to 1.1)
125 µg ethylmercury
602,113
53
0.9
1.0 (0.6 to 1.5)
Autistic spectrum disorders
Vaccination
Person-years at risk
Cases
Cases/10,000
Relative risk (95% CI)
All thiomersal-free
1,660,159
430
2.6
1.0
Any containing thiomersal
1,220,006
321
2.6
1.1 (0.9 to 1.4)
Dose of thiomersal
None
1,660,159
430
2.6
1.0
25 µg ethylmercury
169,920
40
2.4
1.0 (0.7 to 1.4)
75 µg ethylmercury
447,973
130
2.9
1.2 (0.9 to 1.6)
125 µg ethylmercury
602,113
151
2.5
1.1 (0.8 to 1.5)


Comment

What we have here is a superb study of what was, in effect, a real world before-after experiment. The study was huge, and comprehensive, covering almost 99% of children born in Denmark during a period during which a switch was made from use of a vaccine containing thiomersal to one that did not. It was the only vaccine given to children that did contain thiomersal. Moreover, diagnosis of autism or autistic spectrum disorder was according to strict criteria, and comprehensively applied. Follow up was for a minimum of four years, ensuring that almost all cases likely to occur should have occurred during that time.

Of course, Denmark changed to having thiomersal-free vaccines. Even with good evidence of lack of association, that is a good thing. What we have, though, is powerful evidence that autism and autistic spectrum disorders do not arise from use of thiomersal in vaccines.

Reference:

  1. A Hviid et al. Association between thimerosal-containing vaccine and autism. JAMA 2003 290: 1763-1766.

previous or next story