07/10/2013 04:45 EDT | Updated 09/09/2013 05:12 EDT

The Problem With Studying Autism and Diet

A study published in the open access science journal PLoS ONE on July 3rd 2013 is starting to make the rounds in the alternative medicine community. 

One site has already claimed that the new study serves to partially exonerate anti-vaccine dream-date Andrew Wakefield's study on the role of gastro-intestinal (GI) symptoms in autistic children and his now falsified claim that the measles, mumps, and rubella (MMR) vaccine was the cause of not only the GI symptoms but of autism itself. 

Reading these claims, it makes me wonder if any of the Wakefield supporters even read either of the studies, or even the abstracts, because a closer look reveals only the most tenuous of associations, and the retraction by the Lancet of Wakefield's paper is not  in any way in jeopardy.

The new study out of Arizona State University by Kang et al compared the collection of normal bacteria found in the gut of autistic children with the collection, called the intestinal microbiome, in neuro-typical children.  The study of the microbiome and its effect on health has been an exciting area of research over the past several years, with connections between gut bacteria and mental health, heart and stroke, diabetes, and inflammatory bowel disease making their way into the news.  It is certainly known that uncomfortable and distressing GI symptoms have been associated with children diagnosed with Autism Spectrum Disorder (ASD) but the exact mechanism or even a vague notion of how to characterize the symptoms has been elusive.

The team at Arizona State analysed the stool samples of 20 children with ASD and 20 neuro-typical children. The analysis revealed difference in the diversity of specific types of bacteria, with statistically significant fewer bacteria from the genus Prevotella, and to a lesser extent Coprococcus and Veillonellaceae, in autistic children.  That is where most of the reporting ends in the mainstream press, but it is important to insert some caveats before conclusions can be drawn about the significance of the study.  First, the study failed to find any correlation between the severity of the GI symptoms and the severity of ASD, something that was not expected by the researchers.  Even more telling is there was no relationship found between the severity of the GI symptoms and the levels of difference in the microbiome: fewer Prevetolla species did not result in more GI symptoms.   The only correlation that this study found was between the overall decreased Prevotella and the diagnosis of ASD.

When we look at similar studies on the microbiome and ASD, different results have been found as well.  In 2010, a similar analysis did not find the same disparities as the Kang paper.  They found instead an increase in another bacteria species instead.  In a previous paper by one of Kang's co-authors, Adams, they did find a correlation between GI symptoms and severity of ASD.  Because of this heterogeneity in the previous findings, all admitted by Kang et al in the paper, we cannot draw conclusions about what has caused these disparities in bacteria, only that they have occurred in this sample of these particular children.

Though it is technically complex, this study does seem thorourgh, it does not over reach in its conclusions, and it is another step toward understanding the role of the gut in ASD.  What this paper does not have, however, is any relation to the discredited Wakefield study.  Wakefield did not study the microbiome in any detail.  He did look for the evidence of parasites like Shigella, Salmonella and Campylobacter, but he did not assay the biome or look for any abnormalities in it.  There is just no relation to these two studies at all, other than a general interest in the connection between GI symptoms and autism.  Most definitely there is nothing we can conclude about the effects of the MMR vaccine on the microbiome, and anybody who is making these claims is making dangerous speculations

While it is understandable that the public is interested in the ongoing story of why ASD patients have so many problems with their gut and diet, we cannot leap beyond the conclusions from one study and overgeneralise.  This study, like any good science, leaves us with more questions than answers. Is there a connection between neurological symptoms and Prevotella?  Is the decreased incidence of Prevotella predictive of ASD? Is the severity of GI symptoms really connected to the severity of ASD?  Does the ASD cause the change in the microbiome, or does the microbiome cause ASD, or are both conditions interconnected in multiple ways?  Is there a genetic link between the ASD child and the microbiome and how is the microbiome of the child related to his or her immediate family?

What we do know is this will continue to be an active area of study for a long time to come, and Wakefield will still be left out in the cold.