Does autistic enterocolitis exist?BMJ 2010; 340 doi: http://dx.doi.org/10.1136/bmj.c1807 (Published 15 April 2010) Cite this as: BMJ 2010;340:c1807
- Nicholas Wright, warden
- 1Barts and the London School of Medicine and Dentistry, Queen Mary University of London, London E1 2AD
In the linked feature (doi: 10.1136/bmj.c1127),1 Brian Deer tries to unravel the journey of the histopathological reports on the biopsies that formed the basis of the recently retracted study by Andrew Wakefield and colleagues in the Lancet,2 which introduced the entity “autistic enterocolitis.” Apparently, the biopsies were first reported on by Susan Davies, a consultant histopathologist and coauthor on the study, but they were also seen and interpreted by three other coauthors before final publication. One of them was Professor Amar Dhillon, also a consultant histopathologist, although his role was apparently only to grade the biopsy inflammation on a grading sheet.1 However, there has been some dispute about this grading sheet, or proforma. In a letter published in 2007,3 Wakefield stated that it was designed by Dhillon and published by Wakefield’s group in 2000,4 although Dhillon was not a coauthor of the latter paper. This proforma published in 2000 referred to some unusual histopathological changes, such as “disruption of the epithelial basal lamina” and “condensation of the lamina propria”—terms with which few gastrointestinal pathologists would be familiar—and also referred to hyperplasia of mucosal lymphoid follicles, the importance of which is disputed.5 However, in a response to Wakefield’s 2007 letter, Dhillon denied that he had designed the proforma, stating that he provided one that he did design, which was based on British Society of Gastroenterology guidelines, and from which most of these terms are missing.6
The biopsy sections were also examined by Andrew Anthony, then a junior pathologist, and by Wakefield, who as Deer points out is not a professional pathologist but a surgeon. It might be assumed that such a plethora of opinions would have produced consensus about the nature of the histopathological changes. Apparently not. The original hospital pathology reports, implicitly Davies’s reports, described normal appearances in eight of the 11 cases for which reports are available, according to Professor Karel Geboes, a professor of pathology at the University Hospital of KU Leuven in Belgium. But the Lancet study reports “non-specific colitis” in 11 of the 12 children, with varying degrees of activity. When Davies was cross examined before the General Medical Council she said that she had initially been concerned about the use of the term “colitis” in the Lancet paper, which is not surprising, because she herself had found nothing abnormal in the biopsy sections. But she was reassured, she said, by the “formalised review” of the biopsies by her three colleagues. It is not clear whether she was present at the review that changed her opinion so drastically. As reported by Deer, Dhillon has said that he was not involved in applying the term colitis to his histopathological gradings, and that he doesn’t know who did this.
Should we be concerned about this? This question has two facets. Firstly, all of these people were coauthors on the Lancet study and took joint responsibility for its conclusions. It would therefore be assumed that they agreed with its findings. A basic tenet of scientific publication is that, whatever the outcome of a study, the authors must take responsibility at least for the part of the study for which they are directly responsible. According to Deer’s investigation,1 the only person who is said to have seen the biopsy sections (which are apparently no longer available) and to stand unequivocally by the findings is Wakefield, who as Deer points out is not a professional pathologist. Secondly, the original hospital histology reports of normality evolved, through “formalised review” and translation of the proforma, to colitis. Nothing is intrinsically wrong with such a review, but this process should have been controlled, monitored, and led by the professional pathologists in the team. It is by no means clear that this is what happened.
But there is third point. One of the authors of the study (Professor John Walker-Smith) expressed some doubts about the ability of the Royal Free’s hospital pathology service to detect inflammation, and Brian Deer emphasises, quite rightly, that Wakefield was not formally trained as a pathologist.1 But then neither is Ingvar Bjarnason, professor of digestive diseases at King’s College, who is also quoted in Deers’s article. Professor Tom MacDonald is also quoted by Deer, and neither is he a professional pathologist but a gut immunologist at Barts and the London, yet he has written extensively on the histological aspects of this topic (in association with an experienced paediatric gastrointestinal pathologist).
Should we be scandalised by this? Should histopathological interpretation be exclusively left to professional pathologists? The answer to both these questions is no. Many dermatologists are excellent dermatopathologists, although it took some time for the Royal College of Pathologists to recognise this and allow them access to a diploma in dermatopathology, reflecting to some extent the ambivalence of pathologists about clinicians trespassing on their territory. Similarly, colleagues in gastroenterology and hepatology, through long experience, have developed similar expertise in their fields. We should not be suspicious of this—all histopathological interpretation is a matter of opinion—but we should always ask how reliable that opinion is. Naturally, we would give more weight to expert opinion, especially in a field such as paediatric gastroenterology—hence the need for the professional pathologists who were coauthors to have led and taken responsibility for the histopathological interpretation published as part of the Lancet study.
But even so, differences of opinion about histopathological interpretation will always occur. For example, considerable interobserver variability occurs in the interpretation of dysplasia in Barrett’s oesophagus between non-experts and experts, and even between expert pathologists. Similar variability is seen in the interpretation of gastritis, and although pathologists show better agreement at diagnosing inflammatory bowel disease, they tend to overdiagnose the physiological inflammatory infiltrate in colonic biopsies as evidence of colitis and underdiagnose specific aetiological types of colitis.7 This may well have happened post hoc in at least some of the cases described in the Lancet study.
So what are we to make of it all? Twelve years later it is extremely hard to tell. The biopsy sections are not now available for independent assessment, which of course would be optimal,5 although Professor Geboes, quoted in Deer’s article, thinks that three of the Lancet children had focal active colitis. Deer cites one study that reported that none of 49 adults with focal active colitis developed chronic colitis,8 which is understandable, because the most common cause of focal active colitis is infection. But this is not the whole story in children, where focal active colitis can be a more ominous sign. Another study reported that eight of 29 children with focal active colitis later developed Crohn’s disease and one ulcerative colitis.9 But as yet, we have no reliable criteria to identify inflammatory bowel disease associated with focal active colitis. Another problem is that the changes seen in mild focal colitis can be caused by enemas, especially sodium phosphate types. In addition, Professor Bjarnason, also quoted in Deer’s article, is right: an inherent degree of mild inflammation in the caecum, albeit in adults, has been reported in healthy people, and we should therefore be cautious in describing “colitis” that is limited to the caecum as clinically relevant. All these factors underline the difficulty in assessing the importance of any changes and the need for expert opinion.
So, does autistic enterocolitis exist? Children with autism often have chronic diarrhoea, bloating, abdominal pain, distension, and abnormal stool consistency.10 A cross sectional study that compared children with autism with matched controls and children with other developmental disabilities found that 70% of children with autistic spectrum disorder had a history of gastrointestinal problems, compared with 28% of controls (P<0.001) and 42% of those with other developmental disabilities (P=0.03).11 Certainly something seems to be going on. Moreover, several articles describe the association between inflammatory pathology and autistic spectrum disorder 3 Many of these have been heavily criticised5 and defended,3 and the role of chronic constipation and other associated conditions in causing the changes recorded in these patients has been much debated.5 10
Is autistic enterocolitis a histopathological entity or even an entity at all? In view of the lack of data and the entrenched position of many of the protagonists and antagonists, any firm conclusion would be inadvisable. The expert review, referenced by Deer,12 concludes that key areas such as the prevalence and best treatment of gastrointestinal disorders in people with autistic spectrum disorders are incompletely understood, and that evidence based recommendations are not yet available. We should remember, as recent experience in several fields has shown, that although science has its defects, it is a self correcting process. Time is, perhaps, the wisest counsellor of all. In the meantime, this case offers a salutary reminder for researchers and journal editors alike that coauthorship means bearing responsibility for what is written.
Cite this as: BMJ 2010;340:c1807.
Competing interests: The author has completed the Unified Competing Interest form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declares: (1) No financial support for the submitted work from anyone other than their employer; (2) No financial relationships with commercial entities that might have an interest in the submitted work; (3) No spouses, partners, or children with relationships with commercial entities that might have an interest in the submitted work; (4) He has provided expert opinion in the case of Wakefield v GMC and acted as a character witness for Professor John Walker-Smith.
Provenance and peer review: Commissioned; not externally peer reviewed.