Tuesday, March 28, 2006

Autism - It’s Not Always A Natural Variation

Okay, I finally decided to start blogging. I like to think of myself as a moderate in the autism debates, able to see both sides (albeit sometimes with a bit of prodding), and I have learned from people of all opinions. But I also see a discussion that is increasingly polarizing, in which accepting that others have different views is giving way to an advocacy that rejects other viewpoints as unenlightened, wrong, or not yet at their level of "awakening". This used to be a hallmark of the more activist element within the anti-vaccine crowd, but is also becoming increasingly apparent in the “autism is natural variation” crowd too. It is one thing to oppose some of the more contentious attempts to cure autism, e.g. chelation, or lupron (which IMHO is just nuts), or to reject the concept of treatment for oneself. But as this challenge to the more contentious bio-med practices is morphing into a more direct challenge against even mainstream accepted methods of investigating and potentially treating autism, I think it is time for more moderates to join the debate.

So, let’s start out with the basic premise that autism is genetic. An occurrence rate in the general population of 1 in 166 (approx. 0.6%) is widely accepted. Within families in which one sibling is autistic, the odds of another being autistic is approximately 10% (Constantino JN et al, 2006). Among monozygotic (identical) twin pairs the concordance rate for autism ranges from 36% to 91%, with 60% being a widely accepted number (as far as I can tell, e.g. Bailey et al 1995). In addition, one study found that 77% of discordant monozygotic twins had social and communication deficits. Within dyzygotic twin pairs there were some studies showing an increased risk for autism, but other studies have shown only a slight to moderate increase in risk vs. singletons, with ascertainment bias as a potential explanation. So where does this leave us? Basically, since the rate of sibling autism is significantly higher than background rate, this clearly points to a genetic link. In addition, the high concordance among MZ twins vs. DZ twins, and the only slight to moderate and potentially explainable increase in DZ twins over singletons further suggests that genetics, as distinct from prenatal complications of merely being a twin, are a major explanation for the occurrence of autism.

But here’s the issue. If autism is strictly genetic, then why do so many MZ twins of autistics NOT have autism? MZ twins share the same genetic blueprint. Epigenetics, i.e. the regulation of the expression of genes, could be a significant factor, but a recent study published in the Proceedings of the National Academy of Sciences indicates that twins are epigenetically indistinguishable during the early years of life, with differences in their patterns of epigenetic modification developing as they age. So inherited epigenetic differences are likely not great enough to explain why one twin becomes autistic and the other does not.

So, if genetics and inherited epigenetics do not explain why one MZ twin is autistic and the other is not, then presumably the cause is exogenous. To throw some math at this, at a 60% concordance rate of autism, a minimum of 25% of MZ autistic twins needed an additional factor to “cause” their autism. Again, an MZ twin study found that a significant proportion of discordant MZ twins in their sample (77%) had social and communication deficits. But it has been suggested “that in such families a genetic substrate for autism may produce the social and cognitive deficits that are included in the broader phenotype but that an interaction between the genetic substrate and unknown deleterious environmental factors culminates in a 'second hit' that ultimately produces the narrow phenotype for autism.” While one could suggest that the 'second hit' is the result of twinning, the rate of autism in DZ twins vs. singletons suggests that this is not the explanation (if it were then the DZ rate should presumably be significantly higher than singleton rates). And unless one wants to argue that MZ twin autism is different from non-MZ twin autism then the minimum 25% factor would presumably apply to non-MZ twin autistics too.

So, if the logic above is correct then it means that a significant percentage of cases of autism are acquired rather than genetically or epigenetically predetermined or inherited. The 25% is a floor, not a ceiling. The actual number may be 30%, or 50% or 100%. The number itself is not as important as the fact that the logic (assuming it is not fatally flawed, which is a possibility) indicates that an external factor or cause is required in at least some cases, the absence of which would render the person non-autistic.

This logic does not support the conclusion that no autism is purely genetic (75% could still theoretically be purely genetic - or more if the 60% concordance rate is too low). In fact, it supports the conclusion that autism in a significant number of cases involves at a minimum a degree of genetic susceptibility. It does not prove or disprove that the 'second hit' is thimerosal or mercury. It does not argue for or against chelation. It does not suggest that autism originates either in the womb or after birth. It does not suggest or refute the possibility that autism is curable, or that a cure is or is not desirable.

What it does support is the conclusion that at least some autism has a cause other than “natural variation”. If this is the case then investigating and potentially treating autism in those cases in which it is 'caused' rather than inherited is not a priori an ethical violation of one’s natural genetic expression or right to "be" autistic.

More on this to follow.

---

Update - Related posts:

Autism, Genetics, and Evolution (Apr 19, 2006)

Autism and Minicolumns (Sept 5, 2006)

Autism and the Evolution of the Brain (Oct 13, 2006)