We have over the last 10 years evaluated several thousand children on the autistic spectrum who have significant gastrointestinal symptoms. Upper and lower endoscopy and surgical histology have identified mucosal inflammation in excess of 80% of these children. Almost none of these children with biopsy-proven enterocolitis would fit the criteria set out above. Firstly, these children rarely have vomiting, current weight loss (as opposed to failure to gain weight in an age-appropriate manner), or passage of blood per rectum. The requirement is thus narrowed to a child having two of two relevant symptoms – current persistent diarrhea and current abdominal pain according to their criteria, plus a past history of persistent diarrhea excluding current constipation. 

The requirement for the current presence of these symptoms, for 14 or more days continuously, shows a singular lack of understanding of the episodic, fluctuating, and alternating (e.g. diarrhea/constipation) symptom profile experienced by these children. In our experience, ASD children with histologic enterocolitis typically have 1 to 2 unformed stools per day that are very malodorous and usually contain a variety of undigested foodstuffs. This pattern alternates with that of “constipation” in which the unformed stool is passed after many days of no bowel movements at all, and with excessive straining. This group is entirely overlooked by the arbitrary criteria set forth in their paper. With respect to diarrhea and constipation, a detailed discussion of stool pattern in these children is available1 which further highlights the shortcomings of the above criteria. Moreover, the interpretation of pain as a symptom in non-verbal children, as it often manifests as self injury, aggressive outbursts, sleep disturbances, and abnormal posturing, is notoriously difficult. This interpretation requires an insight based upon the correlation of symptoms, histological findings, and response of symptoms to anti-inflammatory treatment. There is no evidence in the Baird et al. paper that these crucial factors were taken into account. This study’s inappropriate symptom criteria would explain the discordance with other reports that have revealed a high prevalence of significant gastrointestinal symptoms in general autism populations2,3.

It is surprising that Dr Peter Sullivan, a co-author on the paper, who presumably provided the above gastroenterological criteria, was not aware of the aforementioned limitations. In his role as a Defendant’s expert in the UK MMR litigation, he will have had access to the clinical records of autistic children with the relevant intestinal symptoms and biopsy-proven intestinal inflammation.

We suggest that the authors might wish to reflect on the ethical implications of setting the bar too high for the investigation of such children by ileo-colonoscopy, with the attendant risk of missing symptomatic, treatable inflammation.

Since the relevant MMR/autism children are considered to be those with regression and significant gastrointestinal symptoms, the appropriate stratification for between-group analyses of measles virus antibody levels has not been conducted; therefore the paper is difficult to interpret, adding little if anything to the issue of causation. Moreover, it is a major error to have presumed that peripheral blood mononuclear cells are a valid ‘proxy’ for gut mucosal lymphoid tissues when searching for persistent viral genetic material.

A further major problem in this study is the number of children who dropped out or who were unable to provide adequate blood samples. We know nothing about either the 735 children who were lost at stage two, or the 100 children for whom blood samples were not available. At the very least, we should be told whether the children who dropped out were likely to be representative of those who stayed in, with regard to the key issues of interest.

For reasons that will emerge in the near future, it would be of interest to know whether siblings of autistic children were included in either of the two control groups. This information is not provided.

As a general observation, this paper contributes nothing to the issue of causation, one way or another. Case definition alone is likely to have obscured the relevant group of autistic children. The study tells us nothing about what actually happened to the children at the time of exposure. We are increasingly persuaded that measuring things in blood many years down the line tells us very little about the initiating events in what is, in effect, a static (non-progressive) encephalopathy unlike, for example, subacute sclerosing panencephalitis, which is a progressive measles encephalopathy. The gut is a different matter, and analysis of mucosal tissues has been very informative, since here, in the relevant children, active ongoing, possibly progressive4, inflammation has been identified.

References.

1. Wakefield AJ. Autistic enterocolitis: is it a histopathological entity? Histopathology 2006;50:380-384.
2.  Valicenti-McDermott M, et al. Frequency of Gastrointestinal Symptoms in Children with Autistic Spectrum Disorders and Association with Family History of Autoimmune Disease. Developmental and Behavioral Pediatrics. 2006;27:128-136.
3. Horvath K, and Perman JA. Autistic disorder and gastrointestinal disease. Current Opinion in Pediatrics.  2002;14:583–587.
4. Balzola F et al. Autistic Enterocolitis in childhood: the early evidence of the later Crohn’s disease in autistic adulthood? Gastroenterology 2007;132:suppl 2, A 660. 

Competing interests
Dr Wakefield, Dr Krigsman, and Dr. Stott acted as Claimant experts in the UK MMR litigation