Sep 13 2010

The Long Awaited CDC Trial on Thimerosal and Autism

We can add one more study to the pile of evidence showing no association between exposure to thimerosal (a mercury-based vaccine preservative) and autism. The article: Prenatal and Infant Exposure to Thimerosal From Vaccines and Immunoglobulins and Risk of Autism, is published in the latest issue of Pediatrics, and shows no association between prenatal and infant exposure to thimerosal and three forms of autism – autism, autism spectrum disorder, and regressive autism.

No one study can ever be definitive, but now we have a large body of evidence from multiple studies showing a lack of association between thimerosal and autism. This won’t stop the dedicated anti-vaccinationists and mercury militia from continuing their anti-vaccine propaganda, but hopefully it will further reassure those who actually care about the science.


This has been a long and complex story, so let me review some of the background. Diagnosis rates of ASD have been climbing for the last 20 years, prompting some to search for an environmental cause. The existing anti-vaccine community, not surprisingly, blamed vaccines. This was given a tremendous boost by the now-discredited study by Andrew Wakefield concerning MMR (which never contained thimerosal) and autism. When the evidence was going against MMR as a cause, attention turned to thimerosal in some vaccines. This notion was popularized by journalist David Kirby in his book, Evidence of Harm.

However, this important premise of a correlation – rising ASD rates – is not as simple as the anti-vaccine crowd assumes. In fact there have been many studies of autism prevalence and the consensus at this time is that most of the increase in ASD is due to a broadened diagnosis, diagnostic substitution, and increased surveillance. There may be a small real increase, but most if not all of the increase is an artifact of diagnosis, not a real increase. (Side note – I have written about this topic many times before, and the links I will provide for background are to my previous reviews and summaries, not the original research. But of course, the links to the original research can be found in my prior articles.)

As further support of this a recent NHS study found a consistent prevalence of autism of about 1% in all age groups. If autism rates were truly increasing we would expect a lower prevalence in older age groups, but that is not what they found.

Another line of evidence is the younger and younger identification of signs of autism. Formal diagnosis is often made around 2-3 years old, after many vaccines are given. But numerous recent studies have documented signs of autism as young as 6 months of age. This makes it difficult to blame vaccines given after 6 months.

Holdouts for vaccines as an important contributor to autism rates have pointed to prenatal vaccines given to the mother, but a study finds no correlation there either.

Where the “mercury militia” gains their only support is from the fact that mercury is indeed a known toxin, and a neurotoxin. However, there are still problems with the notion that mercury toxicity from thimerosal causes any neurological damage or specifically contributes to autism rates. The first is that thimerosal contains ethylmercury, which is not nearly as toxic as methylmercury, the form that is found in fish and other environmental sources. Second, the doses given is vaccines is well  below safety limits. Anti-vaccinationists argue that the cumulative dose is high enough to cause damage, but there is no evidence for this. What there is evidence for is the fact that infants excrete mercury very efficiently, and therefore likely clear their mercury load after one vaccine and before the next, so it does not accumulate. And finally – mercury toxicity does not resemble autism (despite the false claims of the anti-vaccine crowd).

Another line of evidence that presents problems for the vaccine hypothesis of autism is the massive and growing evidence that autism is dominantly a genetic (not environmental) disorder. Of course, genes interact with the environment, and there may be environmental factors, but the dominant factor is genetic.

In addition to the multiple independent lines of evidence all arguing against a link between vaccines in general and thimerosal in particular to ASD, there is the ecological and epidemiological evidence which looks specifically at if there is any correlation between thimerosal exposure and risk of autism. Here the answer is a clear – no. There have been multiple such studies in multiple countries (summarized here) showing no correlation.

In addition, A recent Italian study showed no link between the amount of thimerosal exposure and autism risk. Another study showed no correlation between blood levels of mercury and risk of autism.

But the most compelling evidence against a link came from the removal of thimerosal from the routine childhood vaccine schedule in the US. By the end of 2002 all thimerosal, except for insignificant trace amounts, was removed from all vaccines given routinely to children. Only some flu vaccines still contained a small amount of thimerosal, but this is an optional vaccine with thimerosal-free options. The result was a dramatic plummet of thimerosal exposure in the US childhood population. If thimerosal were a significant contributor to autism incidence then we should have also seen a plummet in autism rates. David Kirby predicted this, and the anti-vaccine movement agreed. They gloated about the day they would be proven undeniably correct.

But now it has been 8 years – and autism rates continue to rise at the same rate, without so much as a blip. They tried to move the goalposts for a while, and make some desperate arguments to rescue their failed predictions, but there is no hope. There is no rational conclusion remaining except that thimerosal in vaccines is not a measurable contributor to autism rates.

The CDC Studies

Three years ago I wrote about a CDC study that looked at thimerosal exposure and 42 different neurological outcomes (but not autism). What they found was that there were a few scores that were worse among those exposed to more thimerosal, but there were also a few scores that were better. There was a random distribution of slight positive and negative effects that essentially average out to no net effect. It’s all just noise.

The bottom line is that this study showed no correlation between thimerosal exposure and adverse neurological outcomes.

At the time we were promised a follow up study of similar design that looked specifically at autism – and now, after a three year wait, we finally have those results. This is a case-control observational study that looked at managed care organization (MCO) members for their history of vaccination, including pre-natal vaccination, as well as exposure to thimerosal through immunoglobulins. They correlated thimerosal exposure from these sources to later diagnosis with autism, autism spectrum disorder, and regressive autism. They found:

RESULTS: There were no findings of increased risk for any of the 3 ASD outcomes. The adjusted odds ratios (95% confidence intervals) for ASD associated with a 2-SD increase in ethylmercury exposure were 1.12 (0.83–1.51) for prenatal exposure, 0.88 (0.62–1.26) for exposure from birth to 1 month, 0.60 (0.36–0.99) for exposure from birth to 7 months, and 0.60 (0.32– 0.97) for exposure from birth to 20 months.
CONCLUSIONS: In our study of MCO members, prenatal and early-life exposure to ethylmercury from thimerosal-containing vaccines and immunoglobulin preparations was not related to increased risk of ASDs.

The strengths of this study are the large numbers, the thorough assessment of ethylmercury exposure, and the confirmation of diagnosis. However, this is not a perfect study – it suffers from the limitations of observations studies, as the authors point out in their discussion. The primary weakness was the fact that of 771 potential case-children and 2760 controls they ended up with 246 cases-children and 752 controls. Around 12% were not eligible for various reasons, and the rest were not able to participate for various reasons (because they could not be located, for example), but most of the drop out was simply because potential subjects did not wish to participate.

This high percentage not participating in the study opens the door wide for bias in the final results. The authors were fairly thorough in exploring possible sources of bias from this fact. They found that study participants did not differ significantly from those not participating in various key aspects – such as having an older sibling with autism. Also most case-children were diagnosed with autism after their infant vaccines, so this unlikely to have affected vaccination rate.

But of course it is always possible for there to be unknown confounding factors biasing the results.  A protective effect from thimerosal is biologically implausible, so these results are due to either random chance or some bias in reporting or participation that is not apparent.

Even still, if there were a significant causal effect from thimerosal it should be apparent in this type of study, and it wasn’t.


No one study, especially an observational study, is ever very compelling. I don’t think this one new study changes the scientific picture of vaccines or thimerosal and autism. But it is one more study that fails to show any correlation between thimerosal exposure and risk of developing autism or ASD. This comes on top of multiple independent lines of evidence all pointing away from the notion that vaccines or thimerosal are a significant cause of autism.

The scientific community is likely to see this as further confirmation of a lack of association between vaccines and autism – just one more piece of the big picture. The anti-vaccine community is likely to dismiss it as either hopelessly flawed or as part of the conspiracy. In other words – this study is unlikely to change anyone’s mind on this issue.

28 responses so far