The final nail in the mercury-autism hypothesis?


Two and a half years ago, very early in the history of this blog, I wrote one of my usual logorrheic (although I prefer the word “comprehensive”) posts entitled Mercury in vaccines as a cause of autism and autism spectrum disorders (ASDs): A failed hypothesis. In that post, I characterized the scientifically discredited notion that the mercury in the thimerosal preservative that used to be in several childhood vaccines was the cause of the “autism epidemic” as “one of the most pernicious medical myths of recent years.” And so it is. I like to characterize the notion that thimerosal-containing vaccines (TCVs) cause autism as the American version of the British myth, popularized by Andrew Wakefield and a sensationalistic British press, that the measles-mumps-rubella (MMR) vaccine causes autism and “autistic enterocolitis.”

Both notions were based on confusing correlation with causation, aided and abetted by some truly bad science, and both notions have been painfully difficult to dislodge. Indeed, in the case of Wakefield, only now that Wakefield was stripped of his license to practice in the U.K. by its General Medical Council, leading to The Lancet finally doing what it should have done six years ago and retracting Wakefield’s 1998 study that sparked the MMR frenzy in the U.K. and arguably kickstarted the modern anti-vaccine movement, do I sense that journalists are finally “getting” that science does not support the idea that the MMR vaccine causes autism. Andrew Wakefield may be trying to fight back with his book Callous Disregard after his disgrace was complete, basking in the glow of admiration of die-hard anti-vaccine groups, but, for now, at least, Wakefield and his MMR fear mongering are yesterday’s news, and that’s a very good thing indeed–at least for as long as it lasts.

Perhaps it is the fall of Andy Wakefield that has led to an apparent resurgence of the concept that mercury in TCVs somehow causes autism, after having faded into the background after the CDC and AAP recommended that thimerosal be removed from all childhood vaccines in 1999 and the last TCV having expired towards the end of 2001. After all, if the hypothesis that TCVs cause autism had been correct, we should have expected to see a marked decrease in the incidence of autism and autism spectrum disorders (ASDs) within about 5 years of 2002, given that the vast majority of cases of ASDs are diagnosed between the ages of 2 and 5. We have not, and, even though its adherents have kept moving the goalposts back regarding the date that we should start to see a leveling off and drop in the incidence of ASDs, starting with 2005, then 2007, and now, apparently, 2011 (which is only less than four months away, by the way), even Jenny McCarthy’s anti-vaccine organization originally founded by J.B. Handley and his wife, namely Generation Rescue, began demphasizing mercury in 2007, after having stated flatly on its website that autism is a “misdiagnosis for mercury poisoning” for so long. Since then, “too many, too soon” has been the favored propaganda talking point.

Of course, not every crank is ready to abandon the myth that TCVs cause autism. Indeed, tomorrow two mercury militia “heavy hitters” and bloggers for the anti-vaccine propaganda blog Age of Autism, Mark Blaxill and Dan Olmsted, will be releasing a book entitled Age of Autism: Mercury, Medicine, and a Manmade Epidemic. In anticipation, four weeks ago I actually e-mailed the publicist to send me a review copy of Age of Autism. I have yet to receive the book. I wonder why. Be that as it may, it amuses me that the official release of the release of the not-so-dynamic duo of the mercury militia’s book actually will one day after a study that is arguably the last nail in the coffin of the very dead hypothesis that TCVs cause autism was released. Either the great pharma conspiracy is far more conniving and effective than even J.B. Handley thinks, or Blaxill and Olmsted’s luck is just that bad. As I anticipate the conspiracy mongering posts about this bad timing aside, let’s just take a look at this last coffin nail, which is a study by Price et al that was released today in the journal Pediatrics entitled Prenatal and Infant Exposure to Thimerosal From Vaccines and Immunoglobulins and Risk of Autism.


The first thing you need to know about this study is that it is the long-awaited follow-up to a study by Thompson et al published in The New England Journal of Medicine in 2007, which I’ve discussed before. Thompson et al was a study that examined in a rigorous fashion whether there is a correlation between TCVs and adverse neuropsychological outcomes other than autism and ASDs. Basically, this was a case-control study of over 1,000 children tested for 42 different neuropsychological outcomes other than autism and ASDs in which children were subjected to various neuropsychological tests and the results correlated to exposure to TCVs during the neonatal period (days 0 to 28) and the first seven months of life. The results were a classic example of in essence random noise. There were a few neuropsychological measures in which lower scores correlated with TCV exposure, and there were a few neurodevelopmental measures in which higher scores correlated with TCV exposure. However, the detected associations were small and almost equally divided between positive and negative effects; in other words, nothing more than would be expected from random noise in the data.

One aspect of this study that was depressing and amusing at the same time was that the input of anti-vaccine–excuse me, “pro-safe vaccine”–groups was solicited during the design of the study. Indeed, Sallie Bernard, president of SafeMinds, was a consultant to the investigators of Thompson et al and participated in the design of the case-control study, despite the fact that she has no relevant background in science, statistics, or clinical trial design! Yes, in the ultimate case of trying to appease those who cannot be appeased and indulging Ms. Bernard’s arrogance of ignorance, apparently the CDC and the Vaccine Safety Datalink (VSD) team decided to “reach” out to their foes. The result? Sallie Bernard saw the writing on the wall, namely that the study was not showing what she wanted it to show, dropped out of the study, and started attacking it immediately, even jumping the gun on the article embargo in order to post her sour grapes all over the Internet. A blogger by the ‘nym of Interverbal aptly criticized Bernard for this:

Even before this study was released yesterday, there was publicized dissent from advocates from the idea of a mercury etiology of neuropsychological harm. One of them came from a well known advocate, Sallie Bernard, who was invited to be a collaborator in this study. It seems that she was involved in the planning of this study, but the lead author indicated that she withdrew her support after the results began to be circulated.

That is not how science works. If you have a problem with a study design then you dissent before you begin collecting data.

As Joseph aptly put it, sour grapes. As the final draft circulated, qq read it and didn’t like what the results showed. She even went so far as to write a letter to the NEJM listing her complaints, to which the Thompson et al responded, skewering all of her objections quite ably. Bernard’s behavior with respect to her involvement in Thompson et al is one reason why I have come to the conclusion that attempting to “build bridges” to the leaders of the anti-vaccine movement is a fruitless and pointless exercise. That is not to say that we shouldn’t try to build bridges to parents who express fears over vaccines because they’ve heard the message of anti-vaccine leaders like Bernard, but trying to convert or coopt high profile anti-vaccine activists like Bernard is a waste of effort.

The other common reaction of the anti-vaccine movement to this study was to cherry pick all the negative outcomes and ignore all the positive outcomes. Steve Novella caught opportunistic anti-vaccine propagandist David Kirby doing just that at the time of the study. Sadly, this dishonesty continues even today. Indeed, Mark Blaxill himself promoted the very same misinterpretation of Thompson et al just last week. Spin and misinformation promoted by the anti-vaccine movement never truly dies. Once a line of attack is developed, no matter how many times it is soundly refuted by science, it inevitably rises again, much like the zombies in any number of movies that I’ve enjoyed through the years. What that means is that defenders of science-based medicine, in order to outrun the zombies when they inevitably rise again, need to do their cardio.

That, of course, and the double-tap. On the other hand, anyone who’s read my stuff for a while knows that for me the double-tap isn’t enough. I like to deal with anti-vaccine pseudoscience using at minimum a .50 caliber machine gun.


When Thompson et al was published three years ago, the authors pointed out that they intentionally did not include autism and ASDs as studied outcomes and that these outcomes would be the topic of a subsequent paper. Price et al is that paper. What one also needs to understand about Price et al and Thompson et al is that these were studies suggested at the time that the CDC and AAP first decided on the precautionary principle to recommend the removal of thimerosal from all childhood vaccines. That it took until 2007 to publish Thompson et al and until now to publish Price et al just goes to show how difficult and time-consuming epidemiological research can be. Finally, given that the data sources and methodology were in essence the same for each of the two studies, we can expect that the anti-vaccine movement will use the same spin and misinformation about Price et al as it did for Thompson et al in order to attack it.

Basically, Price et al is a similar sort of case-control study. A case control study is a type of retrospective trial in which subjects with a certain condition (cases) are matched as closely as possible with subjects without the condition under study (controls), and the two groups compared to look for factors that correlate with the condition. That’s how Thompson et al was performed, and that’s how Price et al was also performed. Being retrospective, such a study can never be as rigorous as a randomized controlled trial or a prospective cohort study. However, given that thimerosal has already been removed from all infant vaccines other than the flu vaccine (and there is a thimerosal-free alternative) and, more importantly, that it would be unethical to conduct a randomized double blind, placebo-controlled clinical trial, this sort of trial is the best evidence that we will be able to come up with.

The specific design of Price et al is summarized in this table (click for a larger version):


Basically, the final two groups that were studied consisted of 256 children with ASD and 752 matched controls. One very interesting aspect that looks as though it were almost certainly placed into the experimental design based on concerns of anti-vaccine advocates like Sallie Bernard is a group of children who underwent regression. Basically, the study examined whether there was a correlation between ASD and TCV exposure. It also examined two subsets of ASD, autistic dsorder (AD) and ASD with regression, looking for any indication whether TCVs were associated with any of them. Regression was defined as:

the subset of case-children with ASD who reported loss of previously acquired language skills after acquisition.

Also, when adding up total thimerosal exposure, the investigators also included any thimerosal exposure that might have come prenatally from maternal receipt of flu vaccines during pregnancy, as well as immunoglobulins, tetanus toxoids, and diphtheria-tetanus. In other words, investigators tried to factor in all the various ideas for how TCVs might contribute to autism when designing this study.

So what did the investigators find? I think you probably know the answer to that question. They found nothing. Nada. Zip. There wasn’t even a hint of a correlation between TCV exposure and either ASD, AD, or ASD with regression:

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.

The last result is a bit of an anomaly in that it implies that exposure to TCVs from birth to 1 month and birth to 7 months actually protects against ASD. The authors quite rightly comment on this result thusly:

In the covariate adjusted models, we found that an increase in ethylmercury exposure in 2 of the 4 exposure time periods evaluated was associated with decreased risk of each of the 3 ASD outcomes. We are not aware of a biological mechanism that would lead to this result.

Of course, two of the most likely explanations for such a paradoxical result would be either that parents of cases, given the genetic component of ASD, might have older children who have already developed ASDs. If these parents have imbibed the anti-vaccine propaganda that is so prevalent, they might be less likely to vaccinate their children according to the recommended schedule. The authors looked for such a correlation between older siblings with ASD and TCV exposure levels and found none. They also asked whether parents of children in the case group may have suspected that their children had an ASD and been influenced in their choice of vaccines by that knowledge, but none of the case children had been diagnosed by 7 months and only a few had been diagnosed by 20 months, which were the two time periods for which cumulative thimerosal exposure was calculated. In light of this, I would tend to interpret this seemingly paradoxical result as meaning, in essence, that there really isn’t even a whiff of a hint of a difference between the two groups.

This study had an additional strength as well, namely that the case and control populations were collected from three managed care organizations (MCOs) that participate in the VSD. Consequently, because of the detailed records maintained by these MCOs, investigators were able to develop a detailed and accurate estimate of total thimerosal exposure from the computerized databases maintained by the MCOs as well as the medical records of the cases, controls, all supplemented by standardized interviews with the parents. In addition, outcomes were measured in clinical settings using standardized assessment tools. In Price et al, the most up-to-date standardized assessment tools used to diagnose ASDs were used to identify cases. In addition, in order to make sure that the controls did not include children with undiagnosed ASD, which would tend to decrease any apparent differences between the groups, controls the lifetime form of the Social Communication Questionnaire was administered as part of the interview with each mother for children who had indications of any neurodevelopmental difficulties. Several children were excluded from the control group in this manner. Finally, the detailed medical records and databases maintained by the MCOs allowed for the detailed determination of and control for many potential confounders. All of these are strengths that were shared with Thompson et al.

So is this study the “last nail in the coffin” of the hypothesis that TCVs cause or contribute to ASDs? Scientifically, I would argue that it’s close, if not actually the final nail. After all, there have been multiple large, well-designed epidemiological studies of varying designs, all of which have come to the same conclusion: There is no detectable correlation between exposure to mercury in vaccines and ASDs. As far as case-control studies go, Price et al is quite good, but it is a retrospective study and the possibility of undetected biases or unidentified confounders can never be completely excluded. Also, it shares one other weakness with Thompson et al, namely a relatively low response rate of around 30%. However, as the authors pointed out in their response to Sallie Bernard, that participation rate was actually higher than predicted as the study was being designed, and it was accounted for in incredible detail in the technical description of how the study was set up.

Even though Price et al provides yet another bit of powerful evidence that, as far as science can tell, mercury in vaccines is not a cause of ASDs, I am under no illusion that this study will put this issue to bed once and for all. After all, if Blaxill and Olmsted are still publishing books based on Olmsted’s laughably poorly researched series Age of Autism. Let’s also not forget that Sallie Bernard was an external consultant for Price et al, as well. No doubt she will make her displeasure known soon, and no doubt it will consist of the same already refuted criticisms that she leveled at Thompson et al.

I can think of one criticism of this study that the anti-vaccine movement will level at it. In the methods, the authors state:

Children were excluded if they had the following medical conditions with known links to ASD traits: fragile X syndrome; tuberous sclerosis; Rett syndrome; congenital rubella syndrome; or Angelman syndrome.

Bernard complained that low birth weight children were excluded from Thompson et al (and presumably also from this study), not understanding that the reason for excluding low birthweight children was obvious: Such children are more likely to have neurodevelopmental problems completely independent from any external cause, such as thimerosal. Including preemies and lower birth weight children would only contribute to the background noise and make finding true associations more difficult. Or perhaps she did understand that and picking up false positives from random noise is what she hoped to see. Be that as it may, the same reasons apply for excluding children with medical conditions with known links to ASD traits. Including them would have added random noise to the current study. Anti-vaccine zealots will no doubt claim that the CDC intentionally excluded them because they are more “vulnerable” to “vaccine injury,” but there is no convincing evidence that this is so, and making such a claim is a shifting of the goalposts from the original claim that autism is a “misdiagnosis for mercury poisoning.” Such a hypothesis will probably eventually have to be studied, but it is a relatively implausible hypothesis given the weight of existing evidence. I also wonder if anti-vaccine activists realize just how much of a shrinking of their hypothesis it is to narrow it down to saying that vaccines cause autism only in children with these conditions. Particularly ironic is congenital rubella syndrome, given that maternal rubella during pregnancy is one of the few known external contributors to the development of autism.


In the end, it is always frustrating to watch how studies like Thompson et al and Price et al are spun by anti-vaccinationists. Epidemiology is like that, though. It’s virtually impossible to conduct a case-control study like this without there being significant shortcomings in it. The reason is that, unlike a bench experiment, the investigators can never control all the variables. Trade-offs are inevitable, and rarely are there adequate resources to assure sample sizes large enough to be completely bullet-proof or to be able to account for every single potentially confounding variable.

However, if there’s one rule in science-based medicine, it’s that no one study is ever sufficient to confirm or rule out correlations between undesirable outcomes and various exposures. However, as the weight of several studies starts to bear down on the problem, the preponderance of evidence must at some point be acknowledged, because we do not have unlimited resources to keep doing studies to answer the same question over and over and over again and every repeated study uses resources that could be used to study other potential causes and treatments for autism. Price et al happens to be one large and convincing chunk of that evidence, but it is not the only one. It builds on multiple other studies and it fits into the confluence of evidence strongly refuting the hypothesis that mercury in vaccines is a cause of autism.

Not that the die-hard cranks will ever accept that result. Blaxill and Olmsted’s book still stands there, thumbing its nose at science.

Posted in: Neuroscience/Mental Health, Vaccines

Leave a Comment (56) ↓