A STATEMENT BY DR SIMON MURCH
These allegations concerning our
1998 study are extremely serious, and clearly
require immediate clarification. I welcome
the opportunity to do so. My comment relates
to the alleged lack of Ethical Practices
Committee approval. I refute the allegation
absolutely on the basis of extensive
documentary evidence.
The protocol for the 1998 Lancet
paper was submitted on September 16,
1996, to what was then termed the Ethical
Practices Sub-Committee. It was entitled
“A new paediatric syndrome: enteritis
and disintegrative disorder following
measles/rubella vaccine”. It was signed
by Andrew Wakefield as lead investigator.
Named consultants were John Walker-Smith and
myself, with signed collaborators Peter
Harvey, for the department of neurology, and
Mark Berelowitz, for the department of child
psychiatry. The application was initiated due
to findings at colonoscopy of two children
with behavioural disorders, which would now
be classified within the autistic spectrum,
and a history of chronic gastrointestinal
symptoms, and recognition of a broadly
similar clinical history among other referred
patients. Specifically, for several years
previously we had looked after an autistic
child with severe ulcerative colitis who
eventually required colectomy (not included
in the study), and the second child
colonoscoped (on September 2, 1996) had
ileitis of sufficient extent that a diagnosis
of probable Crohn’s disease was made.
Following this diagnosis, the child had been
entered in good faith by our inflammatory
bowel diseases fellow into an ongoing
(ethically approved) study of polymeric
enteral nutrition. He had already made
remarkable symptomatic improvement, including
apparent cognitive advance. We, thus,
appeared to be dealing with a condition of
significant severity, and had seen clinical
improvement unprecedented in this
child’s history. News of this
improvement was rapidly disseminated among
parents of autistic children, which I believe
led to many further referrals. This child was
included in the study, with additional
investigations performed after ethics
approval was obtained.
The title of this submitted
application is a point of contention, and
should be clarified. Having taken initial
advice from our psychiatric colleagues on the
basis of referral letters, it was considered
that these children demonstrated a form of
autism called disintegrative disorder
(Heller’s disease). After full
psychiatric assessment of each child seen, it
was later concluded that the more accurate
description for the submitted paper should be
pervasive developmental disorder. Our working
title for these cases had, however, remained
disintegrative disorder, while some parents
referred to their child as autistic, and
others did not. The whole area of
nomenclature in autistic spectrum disorders
was notably difficult at that stage. As we
saw more patients, we moved towards a more
inclusive label of autism, which was used in
subsequent correspondence after February,
1998, to the Ethical Practices Committee.
Measles and rubella were singled out in the
application since these conditions, but not
mumps, had been linked to autism in previous
isolated reports.
This application (172-96) was
for permission for in-depth analysis of 25
patients, referred either by general
practitioners or the vitamin B12 unit at the
Chelsea and Westminster Hospital, who had
been studying B12 absorption in children with
regressive neurological disorders. The
selection criteria explicit in this
application were the presence of
disintegrative disorder, symptoms and signs
suggestive of gastrointestinal disease, and
parental request for investigation. All
patients reported met these criteria. The
consultant paediatricians responsible for the
children’s care decided on the
investigations, although advice was taken
from colleagues at other centres. We
determined that these investigations were
required clinically, not only to characterise
gut inflammation but also to exclude primary
neurological diseases. We had in particular
taken advice for the neurological
investigations, since some of the referrals
appeared to have suffered an encephalitic
illness, and specifically the inclusion of
lumbar puncture was suggested to us as
important for assay of cerebrospinal fluid
lactate, to exclude mitochondrial cytopathies
that can cause both neurological regression
and bowel disease. Several of these cases had
not been investigated to exclude a primary
cause of their regression, and we thought it
important to ensure that we were not missing
underlying metabolic or genetic abnormality.
Proposed investigations thus included
ileocolonoscopy and upper endoscopy, barium
follow-through if ileitis was identified,
lumbar puncture (if sufficient fluid remained
after lactate assay, serology and/or cytokine
testing would be performed), magnetic
resonance imaging of the brain to exclude
structural defects, electroencephalography to
exclude covert epilepsy, electrophysiological
testing, and a panel of standard laboratory
tests, with isolation of DNA for complement
genotyping, since C4 deficiency had been
reported to be an association.
The protocol was referred back
at first submission in November, 1996, with
clarifications and amendments suggested, and
was approved in December, 1996. This protocol
formed the basis for all children
investigated in the 1998 Lancet paper,
and all were investigated. We had no idea at
the time of our Ethical Practices Committee
application that lymphoid hyperplasia would
prove so common, although it was a prominent
part of the final report.
It is important to document
where the protocol differed from the
submission. First, neither I nor my fellow
endoscopist, Mike Thomson, eventually
considered it justified to perform upper
gastrointestinal endoscopy in most
patients—there was then no published
evidence of upper gastrointestinal pathology,
and we were performing these procedures under
sedation, as was then our practice. Getting
the precise level of sedation is not easy in
children with such behavioural difficulties,
and we felt this was not appropriate at that
time, although our policy altered in later
years. Second, in the event, we did not
continue with this extended protocol for the
full 25 patients, again because of the
clinical concerns of myself and my
colleagues, since we had found no evidence of
underlying metabolic abnormality in any case
and did not consider that lumbar puncture of
further cases was indicated. Other children
subsequently seen were thus not subjected to
this extended protocol, and investigated by
testing of inflammatory markers and abdominal
X-ray, with endoscopies performed if thought
clinically indicated, unless there were clear
clinical reasons to perform additional tests.
Following the publication of the
initial report, John Walker-Smith sought
guidance from the Ethical Practices Committee
about further investigation of future cases,
stating “I would like formally to
request Ethical Committee approval for our
clinical research analysis of these children
who we are continuing to see by clinical
need”. In a letter to the ethics
committee, further studies were referred to
under the title “autism and non-specific
colitis and Lymphoid Nodular
Hyperplasia” since that was the clinical
entity that the earlier study had defined.
This was reviewed on July 22, 1998, and data
collection from clinically indicated
investigations was approved. This was for
study of subsequent patients investigated on
the basis of gastrointestinal symptoms and
initial assessment, and in no way relevant to
the 1998 Lancet paper, which had been
conducted entirely according to the 1996
approval. Thus, there was no change in the
name of the ethical approval requested for
the 1998 paper, as mistakenly alleged.
A local review initiated by the
Royal Free medical school in July, 1998,
confirmed that the application had been fully
considered by the ethics committee, and that
assurance had been given that the
investigations were clinically indicated. It
was also apparent that the continuing
investigation of those children had been
reviewed by the ethics committee in July,
1998, and appreciated that investigation of
children seen after publication had become
less extensive, and usually restricted to
gastroenterological testing as thought
clinically appropriate.
We contended then, and still
contend now, that these were standard and
appropriate gastroenterological and
neurological investigations for the symptoms
reported given the current state of knowledge
at that time. Undoubtedly we now perform
endoscopy less frequently, but that is based
on extensive experience. Similarly, a child
with coeliac disease in the 1970s would have
had three diagnostic biopsies compared to the
one, or even none, now performed.
Thus, I can confirm that the
patients presented in the Lancet study
were investigated in accordance with the
ethics committee approval of December, 1996,
and that no attempt was made to seek
retrospective approval.
Dr Simon Murch
Senior Lecturer and Consultant
in Paediatric Gastroenterology, Royal Free
and University College Medical School