Emily Willingham


Stomach Endoscopy. Source: Wikimedia Commons

Are autism, gut inflammation, and immune issues linked? One recent sponsored supplement published in the journal Pediatrics argues that they are. There’s certainly some evidence, although quite mixed, hinting at a link between immune or gut issues and a subset of people with autism, although which one triggers the other is nowhere near resolved. What one might hope is resolved is that it’s difficult to construct solid scientific arguments on a shaky foundation that incorporates retracted papers.

Th Pediatrics GI-autism supplement reflects information derived from a symposium conducted in 2009, one convened to talk about GI problems and autism. The authors are recognizable as those whose interest has very much been focused on autism as inflammation, particularly gut inflammation. The problem has been pinning down specifically what that inflammation might be and where it occurs. To demonstrate that such evidence has been identified, the authors write:

Endoscopic analyses of children with ASD and GI symptoms have revealed the presence of a subtle, diffuse inflammation of the intestinal tract (reviewed in refs 9 and 25). Characterizing the nature of this inflammation remains an area in need of further investigation to fully understand and to provide further evidence of its relationship to GI symptoms in individuals with ASDs.

I’ve followed the studies on GI and autism, in part out of curiosity and in part out of a confirmation bias because our autistic son had gut problems early on (that he no longer has) while our youngest, non-autistic son has persistent GI complaints. My general impression has been that data confirm some heightened gut issues related to constipation and anxiety, although one large, 30-year study found no link at all between GI disease and autism. Some research suggests autism-specific pathways involved in children who do have GI problems. But results from study to study are mixed and yield a fuzzy picture. Even the data on the microbiome — the microbial inhabitants of the intestinal tract whose identity and numbers can influence health — are in conflict, as they are for whether or not specific autistic behaviors or severity can indicate underlying GI problems.

Indeed, study findings show a huge variation in percentages of autistic people with reported GI problems and how they compare to those without autism, as this graph illustrates. Of course, each study given in that graph had a different population with different numbers and probably variable versions of autism, so no one can say firmly what the frequency of GI problems really is among
autistic people.

One impression I had, though, is that there’s a dearth in the scientific findings of intestinal clinical signs, such as from endoscopy, in autistic people, which has made it rather difficult to establish any definitive link. Some clinical findings point to esophageal dysfunction: one study found more reflux in autistic children, another identified a rare esophogeal disorder associated with three cases of autism, and yet another found esophageal dysfunction in girls and women with Rett syndrome, a genetic form of autism. Another option for explaining at least some GI problems in autism is a non-infectious, non-inflammatory cause linked to misfires of the enteric nervous system, the ‘little brain’ that resides in your abdomen as a multiplex of nerves that indeed sometimes seem to have a mind of their own. But confirmatory endoscopic evidence for the lower gut is scant, as far as I knew. So when I read that above quote about endoscopic analyses showing a “subtle, diffuse inflammation” in the intestinal tract, my first thought was, “Eh?” It’s a compelling and confident statement of a link between gut inflammation and autism, but what is the evidence?

So, I looked at references 9 and 25, cited in the quote above. As the parenthetical says, they are reviews. Reference 9 is a consensus statement published in Pediatrics in 2010. I searched it for mentions of inflammation and endoscopy. First, I note that it says the following about GI disorders and autism:

The existence of a gastrointestinal disturbance specific to persons with ASDs (eg, “autistic enterocolitis”) has not been established


The panel concluded that evidence-based recommendations are not yet available. The consensus expert opinion of the panel was that individuals with ASDs deserve the same thoroughness and standard of care in the diagnostic workup and treatment of gastrointestinal concerns as should occur for patients without ASDs.

Referencing the findings of gut inflammation, this 2010 review says:

Researchers in a number of laboratories have described intestinal inflammatory changes in individuals with ASDs

and cites three references, 59, 60, and 61.

Note that this consensus statement was published just before The Lancet retracted the infamous 1998 Wakefield et al. paper that set off the whole treacherous distraction of “vaccines cause autism.” In fact, it cites that paper, although discussing it with a critical eye.

Curious about references 59, 60, and 61 embedded in the 2010 Pediatrics consensus paper, I first checked Reference 59.Here it is.

The senior author on this paper is one Andrew J. Wakefield. The report appeared in 2003, years before the 2010 retraction of Wakefield’s now-notorious 1998 paper. Reference 59 relies on the 1998 retracted paper, citing it four times, and cites another retracted Wakefield research article eight times.

What about Reference 60? It’s a review, also published in 2003. It depends so heavily on the retracted 1998 Wakefield paper that it cites it 10 times and cites another retracted
Wakefield paper from the year 2000 twice.

So, on to Reference 61. It’s an original research paper from 2004. This 2004 article features Paul Ashwood, who makes co-appearances with Wakefield in other publications, including as first author on Reference 59, above. Reference 61 cites a retracted Wakefield paper from 2000, its first-listed reference, four times. Three Reference 61 authors were also authors on both the 1998 and 2000 retracted publications.

That’s the end of the references in this review, one of two cited in the 2012 Pediatrics supplement as evidence of clinical findings of “diffuse inflammation” in the GI tracts of autistic children. So far, one of the two references given to
support clinical evidence of diffuse gut inflammation in autistic children sources three references that rely in part or considerably on retracted papers, and none of the sources notes the retractions or contains an erratum.

The second citation for clinical signs of inflammation from the 2012 Pediatrics supplement on GI and autism? Reference 25. That paper is a chapter in a book that was published in 2010, and the chapter cites and discusses extensively and without critique the retracted Wakefield paper from the year 2000. The lead author on this chapter is Paul Ashwood, who also is an author
on this latest 2012 Pediatrics supplement.

In the end, those two reviews that the 2012 Pediatrics supplement cites to support “diffuse inflammation” trace back to only two relevant original research papers, both of which contain several references to retracted Wakefield papers, and one of which has Wakefield as senior author. These two original articles are a legitimate part of the literature; they have not been retracted. Yet even though they rely on retracted publications for rationales, context, and even methods, not one of them appears to contain a note or erratum. Retracted papers should be treated in the scientific literature as though they no longer exist, except as cautionary tales. What would these two original research articles or the cited reviews look like if all references and text related to these “non-existent” papers were removed?

For what are probably obvious reasons, the authors of the 2012 Pediatrics supplement sidestepped using the name “Wakefield” in their reference list by citing for their endoscopic evidence reviews that in turn cited Wakefield and other papers. But that confident statement about endoscopic evidence of intestinal inflammation is based on publications that carry a certain air of fragility thanks to reliance on references
that, in the understanding of scientific retraction, should simply no longer exist. And I find that less than compelling.

This article was published under a different title at www.forbes.com.