Two Cheers for “Double Hit” Autism Theory

A recent study in the July 4^th Archives of General Psychiatry by Joachim Hallmeyer and colleagues at Stanford University indicates recurrence of strict autism within both identical and fraternal twins is lower than found in previous studies.  This suggests that a larger component of the cause of autism in such individuals may be exposure to something either prenatally or after birth that increases autism risk, but mainly in the more severe forms of autism.  Reports in the popular press would have you believe the study indicates genetics are not a major cause of autism, and some kind of "environmental" factor, is likely responsible for autism.  To put is succinctly, hokum.  That is not what the evidence showed.

The study showed that for Autism Spectrum Disorder (including PDD-NOS, Asperger and higher functioning autism not meeting the more severe cutoff score) recurrence in males was 0.77 for identical and 0.31 for fraternal twins; for female the concordance* was 0.50 for identical pairs and 0.36 for fraternal pairs.  The recurrence of ASD within twins reported here is similar to earlier studies, but somewhat lower, while the findings for “strict autism”(i.e.more severe autism) are not.  For example, Ritvo, Freeman et. al (1985) ) [Am J Psychiatry. Jan;142(1):74]  reported concordance for autism of 95.7% in the identical twins and 23.5% in the fraternal twins, and a more recent study reported concordance was 31% for fraternal and 88% for identical twins. Female and male identical twins were 100% and 86% concordant, i.e. if one had autism so did the other (Rosenberg et.al, [2009] Arch Ped Adolesc Med 163:907)

So the long and short of it is, for people who would like to believe there is a significant non-genetic contribution to autism, they can point to the Hallmeyer findings for the more severe forms of autism and shout “Whooopeee!”   For those who think the primary causes of autism are of genetic origin, they can point to the findings for people with the less severe Autism Spectrum Disorder phenotype, and say, “See, I told you so,” though there is evidence of another factor in autism outcome in this latter group as well. By the way, few experts ever doubted that. 

It’s important to note that the Hallmeyer study provides no evidence concerning what kind of non-genetic factors may be involved.  The vaccine and mercury folks will no doubt find this encouraging for their cause, though there could be other far more plausible explanations.  

It possible the “non-genetic” effect is actually a different type of gene effect called epigenetic.  The best-known case is called imprinting. Angelman syndrome and Prader Willi syndrome can both be caused by the same genetic error on Chromosome 15, and the particular syndrome that will develop depends on whether that genetic error is inherited from the child's mother or from their father.  Such “parent of origin” effects modify the way other genes function. Various chemicals, such as some cancer causing toxins can also cause epigenetic changes leading to bad developmental outcomes.  There is no direct evidence at this time that any of those factors are related to autism, but it is possible. There are laboratory
animal studies that show severely stressful conditions can change the way some genes function, another example of an epigenetic influence.  This is a very plausible idea, but the evidence isn't in yet. 

However, in a companion study to the Hallmeyer article it was found there was increased risk of giving birth to a child with autism among mothers taking SSRI antidepressants (like Prozac or Celexa) during pregnancy. The authors say they controlled for whether the person had a mental health problem or not.  But there are several types of mental health problems that might be expected to yield very different results, e.g. OCD vs. major depression.  Depending of how many of one versus the other "mental health condition" were included, the results could vary. There have been numerous studies of prenatal effects of antidepressants, and this is the first to come up with this finding.  It is widely known that anticonvulsants during pregnancy can lead to adverse developmental outcomes, but this is the first strong evidence linking antidepressants and autism.  So if you are taking antidepressants and pregnant, talk with your doctor before stopping medication or thinking about starting an antidepressant. 

The fact that these two new studies have yielded quite different results from similar types of earlier studies warrants further examination of the study methodologies and samples. For the time being, it’s “Two Cheers for the Double Hit Theory, ” the notion that genes alone do not cause autism, but must be impacted by another factor, such as an environmental condition or an epigenetic condition. 

* Concordance or Recurrence measures how likely it is if one pair of a twin has a condition, the other will as well (blue eyes).