Thimerosal, the molecule that doesn't cause autism.

Thimerosal, the molecule that doesn’t cause autism.

One of the most pernicious medical myths of recent years has been the claim, promulgated by a subgroup of parents of autistic children and facilitated by scientists of dubious repute, that somehow the mercury in the thimerosal (ethyl mercury) preservative used in common childhood vaccines in the U.S. until early 2002 causes autism. Although it had been percolating under the radar of most parents and scientists for several years before, this belief invaded the national zeitgeist in a big way in 2005, beginning with the publication of a book by journalist David Kirby entitled Evidence of Harm: Mercury in Vaccines and the Autism Epidemic: A Medical Controversy. The fires of hysteria were stoked even higher by Robert F. Kennedy, Jr., who published a truly twisted and misleading piece of pseudojournalism and pseudoscience published simultaneously in Rolling Stone and on entitled Deadly Immunity. Relying primarily on quote-mining of the transcripts of both a conference held Atlanta by the CDC to discuss the question of whether autism is related to thimerosal in vaccines and an Institute of Medicine report on vaccines while simultaneously misrepresenting the results of two studies by Verstaeten et al to paint a false picture of a government coverup, RFK Jr. almost single-handedly managed to stoke fears that vaccines were causing an “epidemic of autism.”

I say “almost” single-handedly, because, unfortunately, he had help. Relying on the dubious research of a variety of investigators, such as the father-and-son team of Dr. Mark Geier and David Geier, whose prodigious output of badly designed studies emanating from a lab in their home in suburban Maryland, done using a rubberstamp institutional review board stacked with friends and cronies to approve the studies, and published for the most part in non-peer-reviewed journals, activists loudly insisted that mercury in vaccines was the cause of most autism. Others claiming to demonstrate this link include Boyd Haley, a chemist from the University of Kentucky, and a few other vocal scientists and advocates, who claim that autism is, in essence, mercury poisoning. Facilitating the dissemination of this message were reporters such as David Kirby, activists such as Robert F. Kennedy, Jr., and media personalities such as Don Imus. Indeed, some activists claimed that some vaccines were “poisoning” our children, even going so far as show photos of autistic children with the label “mercury-poisoned” underneath them on placards held aloft at protest rallies. They made quite a splash then, and still do to a lesser extent even today. There’s just one problem.

The scientific data, taken in totality, do not support a link between mercury in vaccines and autism. Today yet another important study by Robert Schechter and Judith Grether was released published in the Archives of General Psychiatry entitled Continuing Increases in Autism Reported to California’s Developmental Services System: Mercury in Retrograde1, that utterly failed to support the hypothesis that mercury in vaccines is an etiological factor in autism. It is yet another nail in the coffin of the medical myth that mercury in vaccines causes autism.

Before I discuss this new study in more detail, a bit of background is in order. In response to the FDA Modernization Act of 1997, prior to the hypothesis that thimerosal might cause autism the US Food and Drug Administration (FDA) compiled a list of vaccines and how much thimerosal they contained. Thimerosal had been commonly used to prevent microbial contamination of vaccines, particularly multidose vials, since the 1930s. It could be reasonably argued that, given the more lax standards of the time, thimerosal had not been adequately tested before use in humans, but decades of use after that had, as far as could be discerned, revealed only occasional skin hypersensitivity reactions due to this component. By 1999, under the recommended schedule of childhood vaccines at the time, concern was expressed that infants, before six months of age, were potentially being exposed to cumulative doses of ethyl mercury that may have exceeded safety standards. It should be noted that these safety standards were based on an indirect surrogate of ethyl mercury, namely methyl mercury, and largely in the absence of any real data. In July 1999, the American Academy of Pediatrics and the U.S. Public Health Service decided, as a precaution, to recommend that thimerosal be removed as soon as possible from childhood vaccines.

It did not take long for this recommendation to be implemented. By March 2001, all vaccines in the recommended infant vaccination schedule were available in forms that had at most traces of thimerosal left over from the manufacturing process. The last lots of childhood vaccines with thimerosal had expiration dates in 2002. Indeed, as Arthur Allen documented in his recent book Vaccine: The Controversial Story of Medicine’s Greatest Lifesaver, a survey of several hundred medical offices in February 2002 conducted by the CDC found that, of the three pediatric vaccines that contained thimerosal in the 1990s, only 2% of vaccine stock still contained thimerosal. Since then, with the exception of the flu vaccine, no childhood vaccine in the U.S. has contained more than trace amounts of thimerosal. There has been considerable debate over whether the decision to remove thimerosal was undertaken too quickly. True, at the time it seemed like a prudent, cautious step. However, the decision had unintended consequences. One was that it resulted in a temporary shortage of childhood vaccines. More importantly, though, it fed the fears of activists that the mercury in vaccines must really be harmful. After all, if it weren’t harmful, why would the AAP and PHS recommend its removal?

Why indeed? The use of this precautionary measure, which to health officials seemed prudent at that time, as justification for attacking the safety of vaccines is as good an example of how no good deed goes unpunished as I’ve ever seen. Many parents, faced with the enormous challenge of raising autistic children, not unreasonably wondered whether there was something wrong with vaccines in the first place.

The second bit of background information that you need to know is that, over the last couple of decades, the incidence of autism and autism spectrum disorder (ASD) has increased markedly to an estimated 1 in 150 children. Robert F. Kennedy, Jr. and others who believe that mercury in vaccines somehow cause autism have referred to this increase as an “autism epidemic” (or, more offensively, as an “autism tsunami“) and frequently claim that there must be an environmental factor that has led to this increase. Because the symptoms of autism, such as cognitive delay and withdrawal from interaction with parents, often manifest themselves between one and three years of age and because this is the age when children receive the bulk of their vaccines, there is a correlation. However, correlation does not necessarily equal causation. It may, but often it does not. Often the correlation is spurious, unrelated, or related to a common factor. More investigation is always required to determine if an apparent correlation is or is not due to causation. In the case of autism, there is good evidence, most recently published by Paul Shattuck2, that increased awareness and diagnostic substitution since the criteria for a diagnosis of ASD were broadened in 1994 account for the apparent increase in diagnoses of autism, as pointed out by Arthur Allen and Roy Grinker.

Regarding the question of vaccines and autism, for ethical reasons we cannot do a double-blind, randomized, control trial of vaccines with and without thimerosal. However, we can do the next best thing, and, indeed, we now have several good studies since 1999 that do just that. Some of these studies are epidemiological; some are ecological. What allows us to use them to reject the hypothesis that mercury in vaccines is an etiological agent that is either associated with or causes autism is a very simple but powerful prediction that the hypothesis makes. Quite simply, if the hypothesis is true and thimerosal-containing vaccines (TCVs) cause autism (or are even merely a significant contributing factor), we would expect that the removal of thimerosal from vaccines would lead to a rapid decrease in autism incidence and prevalence within 2-5 years.

There have now been several studies that examined this very hypothesis in countries that removed thimerosal from their vaccines before the U.S. did. For example Hviid et al3 reported that autism prevalence in Denmark increased from 1991 to 1996 despite the removal of thimerosal from vaccines, while Madsen et al4 looked at the time period from 1971 to 2000 and concluded that autism diagnoses continued to increase after thimerosal was removed from vaccines. Neither study supported a causal link between TCVs and autism, and they were a prominent part of the dataset that was used by the Institute of Medicine to conclude in 2004 that there was no good evidence to support a link between TCVs and autism. A more recent study by Eric Fombonne5 in Montreal examined 27,749 children born from 1987 to 1998 attending 55 different schools. Cumulative thimerosal exposure by age 2 years was calculated for the 1987-1998 birth cohorts. This exposure ranged from 100-125 μg from 1987 to 1991, 200-225 μg from 1992 to 1995, and then none after 1996, which was when thimerosal was completely removed from vaccines in Canada. The result was that autism, ASD, and pervasive developmental disorder diagnoses continued to increase in all periods, demonstrating no relationship between TCVs and autism or ASDs. Even more recently, a large study6 failed to support a relationship between thimerosal and adverse neurodevelopmental outcomes, a result that led one of the investigators in the study, Sallie Bernard, a proponent of the thimerosal hypothesis, to disavow the study in a case of sour grapes, because it did not show what she had hoped that it would show.

We are now nearly six years out from the near-complete removal of thimerosal from vaccines. Other than the flu vaccine, there is no more than trace thimerosal in any childhood vaccine; overall mercury exposure due to vaccines has not been this low in decades. Consequently this hypothesis can now be tested in the United States. In a deliciously ironic twist, Schechter and Grether1 chose to use a source of data that has frequently been widely abused by advocates claiming a link between TCVs and autism to try to show one where there isn’t one as though the conclusions were foreordained. Although it is probably not, it has even been referred to as the “gold standard” of autism epidemiology by none other than David Kirby. Indeed, this is the very same database in which David Kirby predicted that there should be a noticeable decrease in new diagnoses of autism by 2007 if the thimerosal hypothesis is true and then later shifted the goalposts to 2011 when it became apparent that there has been no decrease. This source is the California Department of Developmental Services (CDDS) database. The CDDS administers a statewide system of regional centers and developmental centers designed to serve people who are substantially disabled because of autism, mental retardation, or other developmental disabilities. It maintains an archive file of client developmental evaluation reports on clients enrolled in the system. Among the strengths of the system are that it is a population-based system representing the most populous state in the U.S. Moreover, the client reporting form was consistent throughout the study period, preventing confounders due to changes in reporting. The weaknesses of the CDDS is that its data is derived from an administrative system that was designed to track enrollment and fiscal data and is not as well suited to measuring the occurrence of developmental disabilities in the population. However, with proper statistical analysis, considerable information can still be gleaned from this data for specific birth cohorts.

In order to ask the question of whether autism rates had declined, Schechter and Grether examined data for clients with active status reported from January 1, 1995 to March 31, 2007. Using careful statistical analyses, they used two approaches to measure the occurrence of ASD during this period. The second approach, in which ASD prevalence was determined in the 3 to 5 year old cohort, is perhaps the most informative. It shows a continuing increase in autism prevalence without even a blip or decrease in the rate of increase after 2002. Indeed, showing the skill of some bloggers to analyze the same data, the money figure in the paper (Figure 3) looks almost exactly the same as the graph prepared in early 2007, a continually increasing curve since 1995. This result is not only consistent with multiple other published and unpublished studies, including the aforementioned Danish and Canadian studies7, but it is about as unambiguous evidence as can be obtained from a database like the CDDS database. Indeed, despite the limitations of the use of this database, it is an excellent example of proponents of a “mercury injury” hypothesis of autism being “hoisted by their own petard,” so to speak. Indeed, Eric Fombonne, in a blistering editorial8 that accompanies this study, agrees:

The particular significance of the study by Schechter and Grether is that it relies on the California Department of Developmental Services database, which has been systematically used by proponents of the thimerosal hypothesis to argue that the rising number of children accessing these services— or the “epidemic” of autism— was linked to the increasing exposure to ethylmercury of US children occurring in the 1990s through the changes in the immunization schedule. To the contrary, the data analyzed by Schechter and Grether9 provide a clear and unambiguous test that shows that the expected decline in autism rates following discontinuation of thimerosal in US vaccines did not occur.

Noting that, “with the exception of studies conducted by a single pair of authors” (with uncharacteristic restraint Fombonne does not name whom he obviously meant, namely Mark and David Geier), all studies done have thus far failed to find a link between TCVs and autism, Fombonne continues:

Despite the accumulation of scientific evidence rejecting these 2 hypotheses linking autism to various components of childhood vaccines, these theories and the practices that accompany them have not faded away. Why? How many more negative study results are required for the belief to go away, and how much more spending of public funds on this issue could even be justified?

He then postulates an explanation that I happen to agree with:

Outside academic circles, powerful advocacy groups developed and started to lobby decision makers to influence decisions about which autism research to fund and even how to conduct it. Unaware of scientific studies, or worse, doubtful of their results, bestselling writers, journalists, and politicians were drawn to embrace conspiracy theories that portrayed vaccine manufacturers and the Centers for Disease Control and Prevention as public enemies.15 Law firms saw an opportunity to obtain large financial compensations from the US Vaccine Injury Compensation Court or before local federal courts, the viscous US legal process allowing for fermentation of misconceptions. Exploiting further families’ beliefs and their understandable desire to try everything possible to help their children, charlatans developed alternative (and lucrative) “treatments” for autism, which included chelation therapy, use of a hyperbaric oxygen chamber, and testosterone suppression. All are of unproven efficacy, and many are dangerous.

In other words, it’s all about obtaining compensation for nonexistent “vaccine injury” and “biomedical treatments” for this injury. Never mind that these “treatments” are neither scientifically plausible nor have convincing evidence in the form of well-designed clinical trials to support their efficacy in ameliorating the cognitive delays observed in autistic children. Unfortunately, parents who love their autistic children and desperately want to do something to “make them better” are fertile ground for the blandishments of proponents of these implausible and unproven “therapies.” Some of these treatments, such as chelation therapy, which, it is claimed, will remove the mercury that, according to proponents of the thimerosal hypothesis, is the root cause of autism, have developed into veritable cottage industries that prey on desperate parents. It has even progressed to the point where the Geiers can convince some parents that most autistic children exhibit signs of “precocious puberty” and that the elevated testosterone in such children forms “sheets” that bind mercury and prevent it from being chelated properly. As hard as it is to believe, they then use that claim as a justification for using powerful anti-androgenic drugs such as Lupron on autistic children to treat their autism.

Vaccination is arguably the most effective single public health intervention ever developed. As recently as 50 years ago, for example, our parents and grandparents lived in deathly fear of diseases like polio, which is virtually a thing of the past. Because they are preventative in nature and administered to a very large population of healthy people, vaccines have a very high hurdle to jump as far as safety is concerned, because when an intervention is performed on millions of otherwise healthy people, even a low rate of complications can result in large numbers of injured people. Modern vaccines have achieved that level of safety. Are they completely safe? Nothing in medicine is absolutely, 100% safe. In comparison to the risk of the diseases they prevent and by any reasonable standard, the risks due modern vaccines are extremely low. Moreover, the claims of proponents of an increasingly untenable hypothesis to the contrary, there is no convincing evidence that thimerosal-containing vaccines, or vaccines in general, have anything to do with the etiology of autism. Whatever tiny risk there may be from childhood vaccines, autism and ASDs are not among them. Indeed, even before this study by Schechter and Grether, under the onslaught of studies that all fail to find a link between thimerosal and autism, even David Kirby and those more zealous than him were starting to back away from the hypothesis, invoking hand-waving and vague “environmental toxins” or even going so far as to blame mercury from pollution wafting over from China or, even more ludicrously, mercury from the cremation of bodies with mercury amalgam dental fillings. Meanwhile, in the wake of this study, Mark Blaxill is retreating to saying that “the epidemiological analysis doesn’t prove that thimerosal exposure cannot cause individual cases of autism” and blaming vaccines in general for autism (while also not being able to wait for the embargo to try to put his spin on the matter, by the way).

This study is clearly but one more nail in the coffin of this dying hypothesis. Unfortunately, like Jason in the Friday the 13th movies, the hypothesis that mercury in vaccines is a major cause of autism just refuses to die, no matter how many studies fail to find even a wisp of a link between the two. Just when you think it’s finally, really dead, it has an unpleasant way of being resurrected. That’s why it is not difficult to predict that the usual suspects will refuse to believe it, just as they have refused to believe the studies preceding it.


  1. Schechter R and JK Grether (2008). Continuing Increases in Autism Reported to California’s Developmental Services System. Arch. Gen. Psychiatry 65: 19-24.
  2. Shattuck P (2006). The Contribution of Diagnostic Substitution to the Growing Administrative Prevalence of Autism in US Special Education. Pediatrics 117:1028-1037.
  3. Hviid A, M Stellfeld, J. Wohlfahrt, and M Melbye (2003). Association between thimerosal-containing vaccines and autism. JAMA 290:1763-1766.
  4. Madsen KM, MB Lauritsen, CB Pedersen, P Thorsen, AM Plesner, PH Andersen, PB Mortensen (2003). Thimerosal and the Occurrence of Autism: Negative Ecological Evidence From Danish Population-Based Data. Pediatrics 112:604-6.
  5. Fombonne E, R Zakarian, A Bennett, L Meng, D. McLean-Heywood (2006). Pervasive Developmental Disorders in Montreal, Quebec, Canada: Prevalence and Links With Immunizations. Pediatrics 118:e139-50.
  6. Thompson WW, C Price, B Goodson, DK Shay, P Benson, VL Hinrichsen, E Lewis, E Eriksen, P Ray, SM Marcy, J Dunn, LA Jackson, TA Lieu, S Black, G Stewart, ES Weintraub, RL Davis, F DeStefano; Vaccine Safety Datalink Team (2007). Early Thimerosal Exposure and Neuropsychological Outcomes at 7 to 10 Years. NEJM 357:1281-1292.
  7. Parker SK, B Schwartz, J Todd, and LKPickering (2004). Thimerosal-containing vaccines and autistic spectrum disorder: a critical review of published original data. Pediatrics 114:793-804.
  8. Fombonne E (2008). Thimerosal disappears but autism remains. Arch. Gen. Psychiatry 65: 15-6.


Posted by David Gorski

Dr. Gorski's full information can be found here, along with information for patients. David H. Gorski, MD, PhD, FACS is a surgical oncologist at the Barbara Ann Karmanos Cancer Institute specializing in breast cancer surgery, where he also serves as the American College of Surgeons Committee on Cancer Liaison Physician as well as an Associate Professor of Surgery and member of the faculty of the Graduate Program in Cancer Biology at Wayne State University. If you are a potential patient and found this page through a Google search, please check out Dr. Gorski's biographical information, disclaimers regarding his writings, and notice to patients here.