Saturday, 27 July 2013

(28-07-2013) Moms, Are Your Autoantibodies Causing Autism? - Bus1nessN3wz

Moms, Are Your Autoantibodies Causing Autism? Jul 28th 2013, 04:14

"A child with autism (three years old) po...

A child with autism (three years old) (Photo credit: Wikipedia)

The headlines went big with this one, with most of them latching onto a value of 25% or "a quarter" and blaring that maternal antibodies might be linked to that percentage of autism cases. The stories beneath the headlines are almost breathless in some cases in their excitement about the research, talking about biomarker testing, early identification, and 99%–or even 100%–surety that a woman who tests positive for some combination of these antibodies is going to have an autistic child.

The researchers who did the work, led by Judy Van de Water at the University of California, Davis, even coined a term for the kind of autism they say is linked to these autoantibodies: maternal auto-antibody-related autism, or MAR. Yes. MAR.

Antibodies in general are good things, immune system soldiers that tag invaders for destruction. In a pregnant woman, they can and do cross the placenta and reach the fetus, often for good reason–they do what they're supposed to do and protect against threats. But autoantibodies are a rogue breed, going after the "self" (auto) rather than after the "other" that antibodies are supposed to target. The idea with the maternal autoantibody hypothesis of autism is that they cross to the fetus and interact with brain cells in ways that lead to the condition.

In the current study, Van de Water and her team examined a handful of maternal autoantibodies in the blood of 246 mothers of autistic children and 149 mothers of typically developing children. They then looked at associations of inferred autoantibody combinations and having an autistic child. In their abstract, they state:

Exclusive reactivity to specific antigen combinations was noted in 23% of mothers of ASD children and only 1% of controls.

That's where journalists are getting the "almost a quarter" and, in some cases, the loose rounding to 25%.


What the abstract doesn't say is that some mothers with typically developing children beyond that 1% showed reactivity to specific combinations of antigens, the molecular bits that antibodies target. Luckily, this paper is open access, so you can see the table for yourself here. For reasons that aren't quite clear, the authors have broken the combinations into "significant combinations," "specific combinations of two antigens," and "combinations of three or more antigens." The 23% versus 1% values come only from the last category, but if you look at the "significant combinations" of two antigens, you can see that at least 8% and up to 27% (assuming no overlap) of mothers of typically developing children had a reaction to a combination of two antigens. In addition, 70% had a reaction to any one antigen.

A look at the three-antigen (or more) combinations shows that for 9 of the 12 combinations listed, only 1-2% of the mothers of autistic children had reactivity to the triple combination. Of the remaining three, two are combinations to which ~1% of mothers with typically developing children showed reactivity.

In other words, not only did mothers of typically developing children seem to carry some of these autoantibodies and even carry them in triple combinations, but a good percentage of them carry these autoantibodies at least in pairs, and there doesn't seem to be any predictable additivity from having two or three or five.

And then there's some math that's bothered me about the headlines and some of the quotes in the news articles. Thankfully, Jon Brock, an autism researcher in Australia, nicely sums it up in a comment on Ed Yong's well-done article at The Scientist on this study. In the article, Yong quotes Van de Water as saying that if a test based on these antibodies were positive, "their risk is virtually 100 percent" that they will have an autistic child (the story has since been updated). Brock commented:

Based on current estimates, you'd expect roughly 1% (100) of (10,000) mums to have kids to have autism (ignoring for now the fact that some mums have more than one kid with autism). The study suggests that 23% of mothers of kids with autism have a "risky" combination of antibodies, so we'd expect roughly 23 of these mums to test positive.

The specificity is 99%. So of the remaining 9,900 mums in our sample, we'd expect 1% (99) to test positive, in this case falsely because their child is not autistic.

Altogether, we'd expect 23+99=122 mums to test positive. Of these, only 23 (19%) will actually have an autistic child.

Phew. Much better.

The study authors conclude that:

MAR autoantibodies could represent one mechanism underlying the development of one or more of the ASD core features in a subgroup of cases (italics mine). Moreover, with their exceptionally high specificity, several of the MAR autoantibody profiles could serve as the first true biomarkers of ASD risk.

We will see how that goes. Of note, these references to biomarkers often raise the specter for autistic people of a prenatal test and all that can imply. In this case, the stated goal seems to be a postnatal test that might raise an alert to monitor an infant and seek early intervention for any developmental red flags. But another obvious question might be, What does a woman do if she seeks such a test before or during pregnancy and gets a positive result?

Those are the Biggest Questions, but I have a couple of less Big Questions about this study and the maternal autoantibody hypothesis in general.

First, I can't find that the authors state in the paper the timing of the maternal blood draws, but based on what the methods say, it happened after both mother and child enrolled in the study, so well after birth. That leaves open a question of temporal coincidence of these autoantibodies and fetal development. And there's always the other time-related chicken-egg question of who triggered what in whom.

Second, and probably more significant, is a question nags at me when I read these fetal-exposure-to-autoantibody studies. Once mothers make these autoantibodies, they don't stop. Thus, if the autoantibodies are causative or play a big role in autism risk, I'd expect to see a higher rate of younger autistic siblings of autistic children whose mothers test positive for these autoantibodies or fall in the 99% specificity category. The evidence equation adds up like this: Researchers can show (1) the presence of autoantibodies around birth + (2) having an autistic child + (3) a greater autism risk or rate among younger siblings of that child than already exists for siblings of autistic children. Such results would, I think, help solidify these indicators that for a subset of cases of autism, maternal autoantibodies during fetal development play a role.

Now, if we could just get the researchers not to call it "MAR," that'd be great.