Jan 31 2008

Early Diagnosis of Autism – Implications for the Vaccine Hypothesis

Autism spectrum disorder (ASD) is a neurological disorder that includes a wide range of symptoms including reduced social interactiveness and cognitive impairment. At the mild end of the spectrum it is not clear if it is even appropriate to consider the symptoms a disorder and there are those who prefer to characterize such individuals as part of normal human diversity. But more severe manifestations can be significantly impairing.

In the last decade there has been a public controversy over whether or not vaccines in general, or specifically the mercury-based preservative thimerosal in some vaccines, is a major cause of autism. The scientific evidence is very clear – there is no link between vaccines and autism. In addition, the scientific evidence increasingly points to a host of genetic factors as the primary cause of autism. The vaccine-autism hypothesis is subject to falsification from multiple lines of evidence, but there are two that are dramatic and definitive. The first was the removal of thimerosal from most childhood vaccines by 2002. If thimerosal causes autism and was responsible for an increase in ASD diagnosis, then removing thimerosal from vaccines should reduce ASD diagnosis – it didn’t.

The second line of evidence has to do with the timing of the onset of ASD. If A causes B then A must precede B. If it can be shown that B actually occurs prior to A then that would eliminate A as a cause. This is basic logic, something physicians rely upon routinely to help them evaluate possible causes of symptoms and diseases. One of the pieces of evidence proponents of the vaccine-autism hypothesis point to is the fact that parents often notice the onset of autism symptoms shortly after their children receive their first series of vaccines. The scientific community has regarded this as nothing but a coincidence – a post hoc ergo propter hoc (after this therefore because of this) logical fallacy.

The requirement for autism to follow vaccination gives us another opportunity to definitively falsify the vaccine-autism hypothesis. And now we have some good evidence that autism in many cases does in fact precede vaccination. Sara Webb from the University of Washington’s Autism Center has just published a study looking at head circumference in normal and autistic children. She found that in 60% of the 28 autistic boys studied they had accelerated head growth prior to the onset of symptoms of autism. There was no comparison group, but historical controls were used – meaning that it is already well established from routine measurements that children tend to stay on their percentile curves throughout early development. About this study she is quoted as saying:

“This abnormal or accelerated rate of head circumference growth is a biological marker for autism. It occurs before the onset of behavioral symptoms at 12 months of age such as a child’s failure to respond to their name, a preoccupation with certain objects, not pointing to things, a lack of interest in other people and the absence of babbling.

The study needs to be replicated with larger numbers of subjects and a control group, but these early results are compelling. If true it will allow pediatricians to screen for ASD prior to the onset of noticeable symptoms, pushing the diagnosis for many ASD children back prior to being given vaccines.

Head circumference is also not the only line of evidence that is pushing the diagnosis of autism earlier. Several studies looking at home video of autistic children, children with other disabilities, and developmentally normal children show that children with autism display certain features that can reliably predict they will have autism as early as 12 months old. More subtle features are present as early as 8 months old, but these children as not as reliably distinguished from normal controls.

Even more impressively, a study by Philip Teitelbaum showed that careful analysis of the movements of infants from home video could reliably detect autism as early as 4-6 months, and perhaps even from birth.

Taken together there is copious evidence that autism is a primarily genetic disorder that is present at birth, with subtle but increasing signs that separate ASD children from non-ASD children as they age. These fact are incompatible with the vaccine-autism hypothesis. Defenders of the vaccine-autism claim have no answer to this evidence, except to ignore it when possible and dismiss it without justification when forced to confront it. But denying the early onset of autism is becoming increasingly difficult as we learn more about this condition.

To summarize, the vaccine-autism hypothesis has the virtue of being a scientific hypothesis in that it is amenable to falsification. It has now been falsified by multiple independent lines of evidence:epidemiological, genetic, and now clinical. Clinging to a failed hypothesis far beyond the point of falsification is a hallmark of pseudoscience and ideology, and clearly that is what we are dealing with here.

18 responses so far

18 thoughts on “Early Diagnosis of Autism – Implications for the Vaccine Hypothesis”

  1. Simon says:

    Great that we have anotherstring to our argument but you know what they’ll say:

    “Okay, vaccination may not have caused YOUR kid’s autism but it definitely caused MY kid’s.”

    Kids get autism who have never been vaccinated, so the believers must already accept that there must be other causes as well as their own pet one, so they’ll still demonise vaccines. We may even get “The vaccine would make their autism worse” or maybe “It must have been due to the mother being vaccinated”. These people so desperately want vaccines to be the cause of their problems I can even imagine “When they took thimerosol out of the vaccines they must have disposed of it in our water supplies”. They have already shown themselves to be immune from reason.

  2. Michelle Dawson says:

    While generally agreeing, taking on (not for the first time) assumptions made here about “severity” of autism:

    1. Measured “severity” of autism in childhood has been a poor predictor of adult outcomes. We’ve also presented data showing “severity” does not predict level of intelligence in autistics.

    2. The best adult outcomes reported in the literature still belong to individuals who, as children, met the narrowest, strictest, most “severe” autism criteria ever devised.

    3. The only current major instrument that can yield a classification of “mild/moderate” vs “severe” autism (the difference between the two is one point on a 60-point scale) is not based on current diagnostic criteria and no longer considered an adequate instrument for diagnosis in autism research.

    4. Empirically, autistics who have evident strengths (i.e., perform better in certain areas than nonautistics) are considered more “severe” than autistics who don’t. That is, an autistic can move toward being considered “mild” by losing exactly the kinds of autistic traits and abilities that are associated with very good outcomes in adulthood.

    5. In autism early intervention research, greater measured “severity” has been associated with better short-term outcomes. And with worse short-term outcomes. Depending on the study.

    And so on. Yes, there are a wide variety of eventual outcomes in autism (as is true for the nonautistic population), but this is not necessarily associated with “severity” of autism when “severity” is empirically (rather than subjectively) defined.

  3. jim says:

    I realise there is quite some discussion as to whether rates of autism have actually changed, but would it not be possibe to apply any revised diagnostic measures for Autism to the adult population and see if the rate had actually risen or remained the same?

  4. Yes. I have requested the full paper to review, hopefully for Monday.

  5. _Arthur says:

    Notice that the vaccinations and blood tests took place in Argentina in 2003-2004; since then, thimerosal has been removed from childhood vaccines in America, and possibly, in Argentina too.

    That study would confirm several other studies, in lab animals, chimpanzees and humans. The number I recall off-hand is that “most” of the ethyl-mercury was eliminated within 10 days. I cannot cite the precise study for that number, tho.

  6. daedalus2u says:

    The paper is available online


    The highest blood level reported in this study was 40 nM/L, the next highest was 31 nM/L.

    These mercury levels are tiny compared to what was measured in the cord blood of a 996 children cohort in the Faroe Islands where the interquartile levels were 65 and 200 nM/L. That was most likely methyl mercury from seafood.


    There has been a screening of children for autism in the Faroe Islands which includes the cohort tested for mercury at birth (consecutive births for 22 months March 1, 1986- end of 1987).


    They only list the data by year, but in the two years where the mercury measured cohort was born there were 1404 births with 5 children on the ASD. Two with childhood autism and 3 with Asperger’s, or 1 out of 280. In the 9 years studied (1985-1994) there were 7689 children with 12 childhood autism, 20 Asperger’s, 9 atypical for a total of 41, or 1 out of 187.

    Presumably the 996 member cohort represent a good cross section of the population which has very high exposure to methyl mercury yet shows an autism incidence similar to that reported in less mercury exposed populations.

  7. Pingback: RIT Skeptics
  8. arthurgolden says:

    Concerning this post, elsewhere on the internet, Michelle Dawson just posted:

    “latest example of the legendary skeptic Dr Steven Novella resorting to inaccurate, non-subjective characterizations of autistics”

    “And look, no one was the least concerned that all this
    non-empirical, very subjective information about autism was
    being thrown around by a legendary skeptic.”

    Now that this grave concern has been brought to my attention, I am very concerned about non-empirical, very subjective information about autism was being thrown around by a legendary skeptic – if Michelle Dawson could please clearly explain what she is writing about! But quite frankly, based on the limited information provided by Michelle Dawson about the supposed deficiencies in the post by Dr. Steven Novella, I really am left in the dark, as I have been by a number of other “factual” statements she has recently made. On the other hand, her four year-old article “The Misbehaviour of Behaviourists” provides adequate information – but that article prints out to 30 pages. I think Michelle Dawson may cause more harm than good to her concerns when she makes such harsh criticisms without providing adequate explanation for common people like me to be able to understand. I really want to make things better for human beings with autism and I would be glad to cooperate with any one with the same highly ethical goal. I already told her I am willing to discuss these matters if she sends me a private email to golden@shani.net

  9. Michelle Dawson is a fine advocate for autism and has taken a stand against the vaccine-autism quackery. However, she and I differ in that she advocates the position that autism should not be considered a “disorder” but rather it should be thought of as part of the normal spectrum of human diversity. This position is referred to as “neurodiversity” and in the jargon those people who do not have autism are not “normal” but rather are “neurotypical.” Also, within this viewpoint, there is no such thing as “severity” since it is all just natural variation, not a disorder.

    This is a complex question, one that heavily involves the various definitions of entities in medicine and the meaning of “normal” and “healthy”. Without getting in too deep (which I probably will do at some point in the future when I have time for an in depth entry) I prefer the more conventional view that autism can be meaningfully thought of as ranging from mild to severe, and at some point along that spectrum it is reasonable to consider autism a disorder.

  10. Michelle Dawson says:

    Hi Dr Novella,

    Your description of my position isn’t accurate. I suggest reading my work (rather than what other people write about it), if you are going to write about it. I’m not much of an “advocate” except by necessity; and my work has to pass peer review.

    Nor is your description of “neurodiversity” (a word that appears once in my formal writing, and this is still sitting in press) accurate. “Neurodiversity” is not about autism (it is part of the general idea that disabled people should have human rights), and even informally, I’ve written almost nothing about it.

    I don’t use the word “neurotypical.” You will not find it in my formal or informal writing, except when I have to explain that I do not use the word “neurotypical.”

    For what it’s worth, my position can be compressed into two sentences that I’ve repeated a lot.

    1. Autistics are human beings with human rights; and autistics deserve the recognized standards of science, ethics and advocacy that automatically protect and benefit nonautistics.

    2. Services for autistics, whatever those services may be, should be asked for accurately (including with respect to the existing science), ethically, and respectully.

    Nor have I written anywhere that “there is no such thing as ‘severity’ since it is all just natural variation.” I really doubt that would pass peer review, or even be accepted at IMFAR.

    What I did write is sitting up there in the 2nd comment, and appears to have been ignored in favour of a series of stereotypes and caricatures about what people like me are presumed to believe. But the information in my ignored comment arises from existing instruments (anyone familiar with autism will recognize the CARS, e.g.) and the peer-reviewed literature (with the sole exception of the data we presented at IMFAR 2007)–on the off-chance you might be interested in empirical measurements and evidence re autism “severity.” Unfortunately, that doesn’t seem to be the case.

    Also for what it’s worth, the current data-based consensus (as opposed to popularly repeated cliche) is that “severity” of autism is a relatively poor predictor of later outcomes (see Howlin, 2005, for a major review). I provide some instances of this in my post above. There are others.

  11. Michelle,

    Thanks for clarifying your position, and I am sorry that I mischaracterized it. In my haste I made certain incorrect assumptions and have apparently confused your position with that of others. (Actually looking at the resources I was thinking of I see that is exactly what I did.)

    Getting to your specific points – I actually don’t see their relevance to my original blog entry. I only mention severity in my opening paragraph, by way of introduction, and it is not a premise of the rest of the entry.

    My use of “mild” and “severe” were global and subjective – they were not intended to be empirically defined because it was not relevant to my entry. As such it is trivial so say that “more severe manifestations can be significantly impairing.”

    The fact that severity at one age as measured by a specific metric does not predict later outcome is irrelevant. The fact that “severity” is not one-dimensional and will vary based upon what you choose to measure is certainly interesting, but does not invalidate the notion that some people are disabled by their autism and others are not (however you choose to refer to that fact).

    So I guess I am not sure what your point is or how it is relevant. Are you just objecting to my use of the word “severity” without operationally defining it? If that is the case then I have to disagree with your criticism. Terminology is defined by usage, and these terms (“mild” and “severe”) are often used in a global sense without implying anything empirical. It is common practice (at least in my experience) to qualify the terms when wishing to imply an empirical definition. For example, by saying “severity as defined by…” and then giving the specific scale or measure. Often just the measure is given, so if I want to refer to the NIH stroke scale as a measure of stroke severity I will simply say “patients with a higher NIH stroke scale” for example.

    But if I am writing to the lay public I may simply and adequately write: “patients with more severe strokes may be significantly disabled.” And it will be generally understood that I am using the term “severe” colloquially and not implying any specific metric.

    Do you have an alternate terminology you would favor, or do you not think it is ever legitimate to refer to autism “severity”?

    I would be happy to address any further comments you have.

  12. Michelle Dawson says:

    Hi Dr Novella,

    I provided verifiable information about autism “severity.” I did this (again) to indicate, in case anyone’s interested, that there is research and there are data.

    You’re free to promote subjective impressions and assumptions (you know a “severe” autistic when you see one), which are prone to all the usual biases. But I thought I’d point out that popularly repeated assumptions and cliches about autism “severity” might need examining. I thought this was the kind of blog where allusions to critical thinking and the peer-reviewed literature would at least not be totally unwelcome.

    As I’ve commented here before, “Anyone who, like Dr Novella, uses his authority to divide autism by ‘severity’ (or any other way) should have some scientific basis for doing this.”

    “Severity” of autism, in autism research, is the attempt to quantify the extent to which autistic traits and abilities are obvious, in an individual.

    I am wondering who exactly you would regard as “severely autistic.” Autistics who encounter a lot of difficulty do not necessarily present with strong autistic phenotypes (one way of describing individuals whose autistic traits and abilities are very obvious). Autistics who are not obvious, who present with “weak” phenotypes (e.g., they do not meet full autism criteria) sometimes have the most difficulty, for various reasons. This is a clinical observation (not mine; I’m not a clinician or anything close, but I do work within shouting distance of clinicians) with some support in the literature.

    The most obvious (farthest from being “normal”) autistics include autistic savants, among whom are some of the most successful autistic individuals.

    Also, the extent to which any autistic (or otherwise disabled) person is disabled is not necessarily dependent on our own characteristics (including “severity”). This is reflected here and there in the autism literature (both descriptively and empirically). How we are regarded and treated as a group and as individuals has a major impact on how well we can function and on how good our outcomes are. This should not surprise anyone.

  13. Cyanide says:

    Well, we can guess where some of the lawyers for these groups might start moving towards:

    Autism as another manifestation of Cerebral Palsy, caused by birthing trauma. Look out Obstetricians!!!

    Obstetricians love legal battles. They especially love Cerebral Palsy.

Leave a Reply