Immunization Science

Thimerosal Exposure and Neuropsychological Outcomes

The article

Early thimerosal exposure and neuropsychological outcomes at 7 to 10 years. Thompson WW, Price C, Goodson B, et al. New England Journal of Medicine 2007; 357(13):1281-92.


Thimerosal is a compound that is 49.6% ethyl mercury by weight which has been used as a preservative to prevent contamination of some vaccines in multi-dose vials since the 1930s. There has been no thimerosal preservative in any US-licensed vaccines given to children less than 6 months of age since early 2003.

There is no evidence that thimerosal exposure is associated with the development of autism. In 2004, the Institute of Medicine’s Vaccine Safety Committee reviewed the data concerning vaccines and autism—including thimerosal exposure from vaccines—and concluded that the data favored rejection of a causal association. Subsequent studies have similarly found no association between vaccines and the development of autism.

A 2003 study reviewed data from 3 HMOs and also found no association between thimerosal and autism. However, data from one of the HMOs suggested that thimerosal might have been associated with tics or speech delay. No such association was found at the other two HMOs.

This study looked at 42 neuropsychological outcomes—other than autism—in children who had had varying levels of exposure to thimerosal prior to birth and/or in infancy.

The question

Were increasing levels of exposure to thimerosal—prenatal as well as during the first 7 months of life—related to a greater likelihood of non-autism neuropsychological outcomes at 7 to 10 years of age?

The study

This cohort study of 1,047 children born between 1993 and 1998 was conducted using children identified within the databases of 4 HMOs. Children needed to have been enrolled in the HMO at the time of the assessment and to have been in the HMO during the prenatal and postnatal periods that the investigators used to assess thimerosal exposure.

Children were enrolled in the study on the basis of their exposures to thimerosal before birth (prenatal), as newborns (neonatal), and their first 7 months of life. The researchers looked at exposures to both vaccines and immune globulins (which also contain thimerosal) and estimated the amount of thimerosal the infants received from each specific product, and then summed the total exposure to thimerosal for the different time periods; these were expressed as micrograms of ethyl mercury exposure per Kg of the child’s weight.

Trained evaluators assessed 42 possible neuropsychological outcomes at 7 to 10 years of age. The evaluators were unaware of the children’s exposure to thimerosal or their medical history.

The study also examined healthcare-seeking behavior, socioeconomic factors and other possible confounders that might affect the health and development of the children.

The findings

The developmental outcomes of all these children—irrespective of their exposure to thimerosal—were very similar. The study did find a few significant associations between increasing thimerosal exposures and neuropsychological outcomes. The associations that were observed were almost equally divided between better and poorer performances in the tests.

The relevance/bottom line

The findings from this large comprehensive study do not suggest any causal association between increasing exposure to mercury in thimerosal—prenatally and early in life—and neuropsychological outcomes at 7 to 10 years of age.

NNii’s comment

As was the case with autism and thimerosal, the accumulating evidence shows an absence of an association between thimerosal exposure in infancy and the subsequent development of non-autistic neuropsychological abnormalities. Most of the findings—both positive and negative—were likely due to chance.

One weakness of this study was the high proportion of families that declined to participate, raising the issue that there could have been some potential selection biases. However, they were able to study a large number of children.

Because a (not significant) association was observed in a previous study, the findings of an increased frequency of tics and language delay in subsets of children suggest that these findings should be explored separately to assess whether these were chance findings or not.