Controversial research by Andrew Wakefield in the UK that was published in The Lancet in the February 1998 issue, dubbed the "Wakefield Study", suggested a possible link between autism and the MMR vaccine. The original research has come under criticism, largely due to an alleged conflict of interest on Wakefield's part. In March 2004, almost all of the paper's authors retracted its "interpretation" section, which claimed a potential link between pervasive developmental disorders and "possible environmental triggers".
Critics have claimed that Wakefield's study contains many obvious flaws, including an inability to recognize bias in his sample. In October 2005, a study by the respected Cochrane Library said, on the basis of 31 pieces of research into the possible side effects of MMR, that it found no association between MMR and autism. Several independent groups, including the National Academy of Sciences, have also conducted investigations and concluded that the evidence does not support a link.
One study by Gillberg and Heijbel in 1998 examining the prevalence of autism in children born in Sweden from 1975 to 1984 found no difference in the prevalence of autistic children born before the introduction of the MMR vaccine in Sweden and those born after the vaccine was introduced. Another study, conducted by Madsen and other researchers in 2002, studied all children born in Denmark from January 1991 through December 1998. There were a total of 537,303 children in the study; 440,655 of the children were vaccinated with MMR and 96,648 were not. The researchers did not find a higher risk of autism in the vaccinated than in the unvaccinated group of children.
Research in the U.S. has suggested a similar link between autism and the DPT vaccine, although this is not referenced. However, unlike early claims from Wakefield, it is doubtful that a large majority of autism cases would come from this vaccine. Despite all the evidence to the contrary, controversy surrounding autism and vaccines continues to this day, and many polls, such as the autism coach poll, which involved only 15 respondents, show vaccines as the most popular theory currently on the etiology of autism among parents of autistic children.
In a controversial article in June 2005, Robert F. Kennedy, Jr. described research suggesting that it is not the vaccines themselves, but a mercury-based preservative called thimerosal, used in some vaccine preparations (although not MMR), that may be a cause of autism. Kennedy argues that autism was first observed in children who were born around the time of introduction of thimerosal into mass-produced vaccines, and that the incidence of autism in the United States is well correlated with the amounts of thimerosal children receive during their first two years of life. However, the CDC has described a link between thimerosal and autism as 'unlikely'
In 1999 the Public Health Service (including the CDC, FDA, and NIH) recommended that thimerosal no longer be used in vaccine preparations. And while thimerosal is now utilized in only a very few childhood vaccines, it has not been established that autism rates have dropped significantly. The CDC and some medical organizations have repeatedly asserted that no available evidence supports a causal link between thimerosal and autism. Critics have in turn presented criticism of the CDC analyses that suggest deliberate bias in the CDC research.
For example, an analysis of autism rates by Madsen et al. in Demark noted that the incidence of autism remained fairly constant while thimerosal was being phased out and started to rise beginning in 1991, even after thimerosal was discontinued in 1992. Critics of this analysis point out that the methodology was biased. Critics of the Denmark data also point out a significant increase in autism rates among children whose childhood vaccines contained thimerosal. However, in Madsen et al.'s study, the amount of thimerosal in the vaccinations actually decreased while autism rates increased (specifically, during the period 1961–1970, infants had received a total of 400 µg of thimerosal by the age of 15 months, and during the period 1970–1992, infants had received a total of 250 µg of thimerosal at 10 months of age).
The California Department of Developmental Services (DDS), considered to have the best reporting system for autism in the US , has reported unprecedented decreases in new cases entering the system. The number of new cases went from 734 during the second quarter of 2005 to 678 new cases during the third quarter of 2005, a 7.5% decline in one quarter. Note that the total caseload handled by the state is still increasing, but the recent trend points to decreases in the caseload increase. For example, from the 2nd to the 3rd quarter of 2004, the caseload went from 25,020 to 25,769 (749 new cases). Between the 1st and 2nd quarter of 2004, the caseload increased from 24,297 to 25,020 (723 new cases). These variations have led to speculation that removal of thimerosal from vaccines in California is starting to pay off. Others point out they are unremarkable and may simply be an indication that the awareness curve is starting to level off, and that new cases should be expected to decrease to population growth levels (1.6% annual) eventually, probably within a decade or two. Caseload increase between 2004 and 2005 was about 10%.