Pathology reports solve “new bowel disease” riddle

A personal statement by Amar Dhillon in response to “Pathology reports solve “new bowel disease” riddle” (BMJ 2011;343:bmj.d6823). This statement is not a formal representation of UCL’s position.

The reappearance (BMJ Nov 2011 online: http://www.bmj.com/highwire/filestream/536428/field_highwire_adjunct_fil... ) of some of my histology grading sheets for the Lancet 1998 study (Wakefield AJ et al. Ileal lymphoid nodular hyperplasia, non-specific colitis and pervasive developmental disorder in children. Lancet 1998;351:637-41) is interesting. I have not seen the grading sheets since 1997-98 when I gave them to Andy Wakefield. Following the interest shown in the grading sheets in the November 12 2011 BMJ Feature “Pathology reports solve “new bowel disease” riddle” (BMJ 2011;343:bmj.d6823), accompanying articles and editorial it is evident that there are a number of misunderstandings. Many of these are a result of a lack of understanding of the essential difference between the systematic documentation of specific microscopical features in a grading sheet by a “blinded” (ie in the absence of any clinical, or other information) pathologist on the one hand; and on the other hand concluding an overall clinicopathological diagnosis by integration of clinical information with diverse lines of investigation (including information in the grading sheet). The difference between the two activities should be understood better.

The online grading sheets represent an incomplete record of my observations of the slides of gut mucosal biopsies from patients who were included in the Lancet 1998 study, and there were “normal control” biopsies as well; however:
-Of those grading sheets attributed to me not all are mine (grading sheets on p38-47 and p55-64 inclusive of the BMJ Nov 2011 online document belong to someone else: NB the handwriting is different to mine)
-The boxes with assigned patient “case numbers” on the grading sheets have been put onto the grading sheets by someone else (ie “child 1”, “child 2” etc): this information was not available to me at time either of the slide review or Lancet 1998 publication
-Neither the clinical details per case, nor the original diagnostic histopathology reports were available to me at the time of my review of slides
-At the time of my review I had been told that slides of study cases were mixed with “normal” controls: which slides were of study cases and which were controls was unknown to me.

My research review of the slides in 1998 has important differences with the routine diagnostic histopathology process:
-Routine diagnostic histopathology is done with knowledge of individual patient’s clinical details as far as they are available at the time of diagnostic reporting, and so the rendering of a diagnostic histopathological opinion in this situation is usual and appropriate (in direct contrast to the situation that pertains in a blinded research review)
-Then there is a joint review by clinicians and pathologists to evaluate the significance of the microscopic observations in the light of additional clinical, endoscopic, radiologic, and laboratory data that has been obtained after the “diagnostic” biopsy has been reported
-It is not unusual for the clinical significance of microscopic observations to be reinterpreted and altered by this process, and it could be that the histological diagnostic interpretation subsequently has to be corrected.

Thus the purpose of my grading sheet observations in 1998 was not, could not have been, nor was it intended to conclude the final diagnostic assignment of colitis (which has to be made in the light of full clinical/endoscopic/radiologic/ laboratory data; and response to treatment)
-Therefore on the grading sheets “nonspecific” means: “this microscopical appearance doesn’t remind me of any particular disease entity”, and this is why in none of my grading sheet observations have I stated “colitis”.

Bowel disease is not diagnosed by gut mucosal histopathology in isolation:
-I am of the opinion that the histological interpretation should never (or not very often) replace clinical judgement
-“A final diagnosis can only be made with the full clinical information and a biopsy specimen should be reported as diagnostic only if full supportive clinical information is available.” (Jenkins D et al. Guidelines for the initial biopsy diagnosis of suspected chronic idiopathic inflammatory bowel disease. The British Society of Gastroenterology Initiative. J Clin Pathol 50,93-105;1997).

In 1998 there was no paediatric gastrointestinal pathology literature to refer to for guidance regarding ileal and colonic mucosal biopsy microscopical appearances and their interpretation or significance in autistic children. In 1998 the series of cases in the Lancet paper was unusual, if not unique, and it was one of the aims of the study to explore the significance or otherwise of the subtle histological changes in autistic children with gastrointestinal symptoms. Prejudgment of the significance or otherwise of the histological changes in isolation in the 1998 study cases would have been inappropriate previously, and remains so now.
-Several expert gastrointestinal pathologists and gastroenterologists have commented on the grading sheets (BMJ Nov 12 2011) and they have stated that the findings cannot be colitis; however:
-It is a mistake to apply uncritically adult gastrointestinal biopsy histopathological thresholds of normality vs abnormality to children
-The expert gastrointestinal pathologist and gastroenterologist commentators have tried to assess the diagnostic implications of data represented in histopathological grading sheets alone
-This is a fundamental mistake: the significance of the histopathological component of any diagnostic equation depends on consideration of the histopathology within the complete clinical context
- The current opinions of the experts regarding the significance of the histology grading sheets are subject to retrospective bias by knowledge of events since 1998.

At the time of submission of the Lancet 1998 publication I had the clinical, laboratory, endoscopic and histology information presented to me in summary tabular form, and aggregated descriptive text only.
-My grading sheets were with Andy Wakefield and my general recollection of my impression of my slide review was that some biopsies were a bit inflamed, and others were not: I did not know which case was represented by which set of slides, and which sets of slides were “normal” controls. As far as I recall, the changes were not severe in any of the slides, but it is not unusual for gut mucosal biopsies to show little abnormality even in clinically well defined cases of gastrointestinal disease, particularly in children
-My clinical colleagues had collated all of the available information, including my microscopical grading sheet observations in the context of their knowledge of each patient’s condition and concluded a final diagnosis of colitis when this was considered by them to be appropriate
-Thus, at the time of submission of the Lancet 1998 publication, with the limited supplementary information available to me (which I had been prevented deliberately from knowing during the study); and in the context of a comprehensive clinicopathological review by trusted clinical colleagues, the designated diagnosis of colitis seemed to me to be plausible.