 |
|
 |
Autism
and Measles-Mumps-Rubella (MMR) Vaccination: A Challenge
for Pharmacoepidemiology
Posted 12/19/2003
David C. G. Skegg,
M.B., D.Phil. |
|
"Autism
Rocks" T-shirts!!
Note Cards by
people with autism |
Childhood
autism was first described 60 years ago,[1] but
this distressing condition has attracted increasing attention
recently. It is a pervasive developmental disorder defined
by the presence of abnormal or impaired development appearing
before the age of 3 years and by a characteristic type of
abnormal functioning in the three areas of social interaction,
communication, and restricted, repetitive behavior.[2,
3] The parents of autistic children usually recognize
that their child has been unusual since birth or early life,
but sometimes they describe a setback or regression after
a period of apparently normal development.[4] Today
it is common to refer to an autistic spectrum of disorders,
which includes related conditions such as Asperger's syndrome.[4]
Although
the causes of autism are not understood, there is strong evidence
for a neurobiologic basis with an important genetic component.[5]
In a British twin study, 60% of monozygotic pairs, but no
dizygotic pairs, were concordant for autism; 92% of monozygotic
pairs were concordant for a broader spectrum of related cognitive
or social abnormalities, compared with only 10% of dizygotic
pairs.[6] Epidemiologic surveys show a higher prevalence
of autism in boys, and an association with intellectual disability
in many cases.[7] The reported prevalence is highest
in the most recent surveys, but this could reflect changes
in diagnostic criteria and in recognition of the autistic
spectrum (as awareness has increased and services have improved).[7,
8]
Autism
became a center of controversy in 1998, when an article in
The Lancet ignited concerns about vaccine safety. The
authors described 12 children who had chronic enterocolitis
associated with a regressive developmental disorder.[9]
After a period of apparent normality, these children were
said to have lost acquired skills, including communication.
In nine patients the behavioral diagnosis was autism. The
authors suggested that they were describing a unique disease
process, which they later called autistic enterocolitis.[10]
By 2001 they claimed to have found similar colonic lesions
in over 150 children with autism, in whom the main gastrointestinal
presentation was abdominal pain with either constipation or
diarrhea.[10]
The
most worrisome feature of The Lancet article was a
suggestion that the autistic syndrome was precipitated by
measles-mumps-rubella (MMR) vaccination.[9] This
conclusion was based mainly on the fact that, in eight patients,
the onset of behavioral problems was linked, by either the
parents or the child's physician, with MMR vaccination. For
these eight children the average time from vaccination to
first behavioral symptoms was reported to be 6.3 days. The
article was published as an "early report" and was accompanied
by a critical commentary.[11] Nevertheless, it
caused great consternation among parents and some health professionals.
Although the authors acknowledged in their article that they
had not proved an association between MMR vaccination and
the proposed syndrome,[9] there were adverse comments
about the safety of the MMR vaccine at the press conference
that launched the publication.[12]
Few
issues in health care are as emotive as the safety of vaccines
to be administered to children. Even when official bodies
offer reassuring advice,[13] there is a common
tendency to suspect a cover-up. The Lancet authors'
observations were highlighted in the media in many countries,
including the United States. The media attention was most
frenetic in the United Kingdom, where an earlier controversy
about the safety of pertussis vaccination interrupted a successful
vaccination program and led to epidemics of whooping cough.[14]
Autism
tends to appear at around the age when children receive their
vaccinations. The onset is usually gradual, so it is difficult
to see how parents could reliably link the start of the condition
with an event such as administration of a vaccine. Several
correspondents pointed out that the uncontrolled series of
clinical cases discussed in The Lancet article had
not established a temporal association with MMR vaccination,
let alone a causal relationship.[15]
In
a subsequent letter to The Lancet, the lead author
marshaled different evidence in support of his hypothesis.
He presented a graph showing trends in recorded autism in
California and London, England, and concluded that in both
places a marked rise had occurred in the years after MMR vaccination
was introduced.[16] The letter was actually a response
to a detailed analysis of trends in autism in relation to
MMR vaccination in London; the authors concluded that an apparent
increase in the incidence of autism in successive birth cohorts
could not be related to the introduction of MMR vaccination
or to vaccine coverage.[17] Subsequently data from
the General Practice Research Database (GPRD) were used to
address the same question for children from the United Kingdom.[18]
The researchers found that the incidence of autism among boys
(as recorded by family physicians) had increased rapidly during
years when the prevalence of MMR vaccination had been virtually
constant. In another study, the striking increase in the caseload
of children with autism in successive birth cohorts in California
could not be explained by trends in MMR vaccination.[19]
Although
these studies provided strong evidence against the notion
that the increasing rates of autism were due to MMR vaccination,
they did not exclude the possibility that MMR vaccination
could cause autism. This hypothesis was tested directly in
a study from Denmark. Researchers used that country's remarkable
record keeping to link vaccination records with records of
autism. The retrospective cohort study examined data from
more than one-half million children born in Denmark from January
1991-December 1998.[20] Using the unique identification
number assigned to each infant, they linked records of MMR
vaccination status (from the National Board of Health) with
records of autism (from the Danish Psychiatric Central Register,
which contains information on all diagnoses in outpatient
clinics as well as psychiatric hospitals). Information about
potential confounding factors was obtained from three other
registers. The adjusted relative risk of autism among children
who had received the MMR vaccine was 0.92 (95% confidence
interval 0.68-1.24). There was also no increase in the risk
of other autistic-spectrum disorders. This study was highlighted
in a recent systematic review, which concluded that no evidence
of an association exists between MMR vaccination and autism.[21]
The
feasibility of using record linkage for pharmacoepidemiologic
studies in the British National Health Service was also recognized
many years ago.[22] This potential has been realized
through the development of computerized databases such as
the GPRD.[23] In addition to the investigation
described earlier,[18] the Boston Collaborative
Drug Surveillance Program has used the GPRD in other studies
conducted in response to The Lancet article's claims.
In a nested case-control study, results showed no evidence
that children with autism were more likely than other children
to have defined gastrointestinal disorders before their diagnosis
of autism.[24]
In
this issue of Pharmacotherapy, Drs. Jick and Kaye complete
additional parts of the jigsaw puzzle. As well as providing
a new case-control evaluation, they show that the marked increase
in the incidence rate of recorded autism among boys in British
general practices from 1992-2000 was balanced closely by a
decline in the rate of certain developmental disorders without
a diagnosis of autism.[25] This strongly suggests
that the apparent increase in the rate of autism is due primarily
to changes in diagnostic practices and ascertainment. A new
study from California points to the same conclusion.[26]
Five
years after the controversial report in The Lancet,[9]
parents and health professionals can have considerable confidence
in the safety of the MMR vaccine. If vaccination did occasionally
cause an unusual variant of autism, this occurrence would
have to be so rare as to escape detection in well-designed
epidemiologic studies. Moreover, the recent "epidemic" of
diagnosed autism cannot be attributed to the MMR vaccine.
Given that the controversial hypothesis proved to be an expensive
false alarm, should the paper have been published at all?
Although one can only condemn the media frenzy and excessive
claims that stemmed from the publication, this question has
to be answered in the affirmative.
The
reason for this conclusion is that many adverse effects of
drugs and vaccines are first recognized by astute clinicians.
The process of discovery was analyzed for 18 important adverse
reactions that were identified in the two decades after the
thalidomide disaster.[27] In 13 of the 18 situations,
the first alerts came from anecdotal reports, including single
case reports in four instances. Physicians often report their
suspicions in letters to medical journals, but national registries
also are set up to receive reports of adverse reactions.[28,
29] Once an adverse effect is suspected, epidemiologic
studies generally are required to confirm (or refute) the
hypothesis, to estimate the magnitude of risk, and to identify
any other factors that modify the effect.
If
suspicions are not reported promptly, there may be an unacceptable
delay in discovering hazards. Hence it was appropriate for
the controversial hypothesis to be published. Unfortunately,
its reporting in a full article in The Lancet, even
though this was designated as an "early report," gave the
impression that the hypothesis was based on more than a small
series of anecdotal cases and the "hunches" of a few parents
and physicians. Moreover, some propo-nents of the hypothesis
have been reluctant to accept the reassuring findings of subsequent
epidemiologic research, preferring to fuel the controversy
through various channels including the news media.
The
saga about MMR vaccination and autism illustrates the challenge
to pharmacoepidemiologists to be able to investigate suspicions
about the safety of drugs or vaccines in a timely and effective
manner. There can be no doubt that studies using databases
such as the GPRD, as exemplified in the article in this issue
of Pharmacotherapy,[25] have a vital role
to play in this task.
Reprint Address
Address reprint requests to Professor David Skegg, Department
of Preventive and Social Medicine, University of Otago, P.O.
Box 913, Dunedin, New Zealand; e-mail: david.skegg@stonebow.otago.ac.nz.
|
|
|
