Measles-containing vaccines and IBD.
Andrew Wakefield, 02 January 2002
In an ironic twist, investigators from the Centers for Disease Control and
Prevention (CDC) and the Vaccine Safety Datalink (VSD) Project have
confirmed an association between measles-containing vaccines (MCV) and
inflammatory bowel disease. Specifically, they have also determined how age
at exposure to an MCV may be important in determining the type of
inflammatory bowel disease - Crohn's disease or ulcerative colitis - that
Before the American Society of Microbiology's millennium meeting, Davis et
al. reported a retrospective case-control study in which cases (n = 142)
with either Crohn's disease, or ulcerative colitis, were compared with
unaffected controls (n = 432). The exposure of interest was to an MCV.
When performing the power calculations to determine the necessary size of
the study, Davis et al. assumed an MCV-exposure rate of 70%. In actual
fact, the exposure rate was 93-94%, revealing a fundamental flaw in the
study - overmatching.
Exposure of cases and controls to an MCV was so similar that their study,
as designed, would never have detected a risk for inflammatory bowel
disease, even if this risk were real.
Although these crucial data were presented to the American Society for
Microbiology (at the Interscience Conference on Antimicrobial Agents and
Chemotherapy in November 2000), they received no comment in the subsequent
paper published in Arch Pediatr Adolsce Med. An explanation is clearly
In order to determine whether exposure to an MCV is or is not a risk for
inflammatory bowel disease, at the level identified previously , power
calculations reveal that they would have required at least 3 times as many
cases and controls as were included in the study.
The correct power calculations were readily performed using the Epi.Info
package, downloaded from the CDC website.
Age at exposure and risk
In a recent Israeli study, exposure to measles virus was shown to be
associated with an increased risk for Crohn's disease . In addition,
measles virus (including vaccine-strain virus) has been amplified and
sequenced from intestinal tissues of children with inflammatory bowel
disease . However, in the etiology of inflammatory bowel disease, it is
likely that it is the pattern of exposure to measles that is more important
than exposure per se.
Of interest in the Davis study, therefore, was the observation of an
association between age at exposure to an MCV (excluding MMR) and the type
of inflammatory bowel disease that develops . This observation is
entirely consistent with previously reported data, and supports a role for
atypical measles virus infection in the etiology of inflammatory bowel
disease. Such a claim should not be made lightly, and therefore it is worth
reviewing the background to this thesis, that has been described in detail
Age at exposure to natural measles infection is an important determinant of
both acute and delayed outcome.
Early measles infection has been reported as a risk factor for delayed
disease, including subacute sclerosing panencephalitis (SSPE) and IBD
-. For IBD, however, the situation is more complex still; the risk
associated with measles virus includes, for example, concurrent exposure to
measles and mumps infections in childhood  .
Studies have used the UK's birth cohort data (British Birth Cohort study
1970 [BCS 70] and the National Child Development Study [NCDS]), two
population-based, nationally representative, and powerful epidemiological
instruments. These have shown that, within an early window of risk (measles
exposure < 6 years of age) , other characteristics of the acute measles
exposure determine both the risk and phenotype of IBD that develops.
Younger age at infection (high birth order, e.g. two or more older
siblings) (Fig 1) and younger age at exposure to concurrent measles and
mumps infection (Fig. 2) are a risk for ulcerative colitis.
Within this early window of risk, older age at infection (low birth order,
e.g. no older siblings) (Fig. 1) is associated with an increased risk of
Crohn's disease. Higher mean age at concurrent measles and mumps exposure
(Fig. 2), and higher age at monovalent measles vaccination (Fig. 3) are
also associated with an increased risk of Crohn's .
In the BCS 70 cohort, the age of concurrent measles and mumps infection
appeared to determine the risk of whether Crohn's disease or ulcerative
colitis developed in an individual. For Crohn's disease, the mean age of
exposure was greater than for ulcerative colitis. This difference was
statistically significant (p = 0.038).
Data for the birth order/family structure versus risk of Crohn's disease or
ulcerative colitis were then examined to see if this influenced the age of
concurrent measles/mumps infection, as predicted by our hypothesis. This
states that those with no older siblings (low birth order) should
experience concurrent infections later, whereas those with two or more
older siblings (high birth order) should experience the concurrent
The data shown above in Figure 4 indicate that the differential risk for
concurrent infection according to number of older siblings is statistically
significant (p<0.001). Data are adjusted for sex and socio-economic status.
The next question that we addressed was the issue of whether age of
exposure to monovalent measles vaccine (the only measles vaccine available
in the 1970s in the UK) influenced the risk of IBD in the BCS 70 cohort.
The hypothesis predicted that older age at measles vaccination, within the
early window of exposure (0-6 years), would be a risk for Crohn's disease.
This is indicated by the data shown in Figure 5.
In this cohort, age at exposure to monovalent measles vaccine, including
and excluding those with concurrent measles and mumps infections, appears
to be associated with risk of Crohn's disease. This was when adjusted for
sex, social class of father at birth, and crowding. Join the debate! Click
here to post your comments about this Soapbox Speech.
Children exposed the vaccine at an older age are at greater risk (p for
trend = 0.017). This is consistent with the fact the response to the
vaccine virus is a function of age and, possibly, influences of maternal
The numbers of cases are small, and therefore independent corroboration was
required. This was provided by the Davis study, as shown below.
The Davis study identifies that those exposed to an MCV (excluding MMR) at
a younger age (< 12 months) appear to be protected against Crohn's disease,
but at excess risk of ulcerative colitis. Conversely, older age at exposure
is a risk for Crohn's disease, but protective against ulcerative colitis.
This is exactly as predicted by the previous data and the a priori
Analysis of the raw data is required in order to establish whether or not
these trends are statistically significant.
The effects of the characteristics of the exposure are likely to reflect
age- and dose-dependent influences upon immune responses to the virus that
are consistent with models of immunological tolerance, as proposed
The study of Davis et al. provides valuable independent confirmation of a
possible association between age at measles exposure and the type of IBD
What needs to be explained, beyond these observations, is the current
upsurge of IBD in children, including Crohn's disease, ulcerative colitis,
and the recently described autistic enterocolitis.
Were the power calculations for the study, as reported , based upon an
assumed MCV exposure of 70% in cases and controls, as presented in the
original data set?
If the answer is 'no', please would Davis et al. provide the revised
assumptions and power calculations.
If the answer is 'yes', then having performed power calculations for the
size of the study based upon an estimated vaccine uptake of 70%, and
subsequently finding an actual uptake of 93-94%, what steps, if any, were
taken to reconcile this discrepancy prior to proceeding with the study?
Why were the power calculations, as presented to the American Society for
Microbiology, not presented in the paper?
If the answer to question 1 is 'yes', and if no steps were taken to
reconcile the discrepancy, why was this not described in the paper as
If the answer to question 1 is 'yes', then, based upon the obvious flaw of
overmatching, do the authors still consider the conclusions to be valid?
Davis RL, Kramarz P, Bohlke K et al.. Measles-mumps-rubella and other
measles containing vaccines do not increase the risk for inflammatory bowel
disease: a case-control study from the Vaccine Safety Datalink Project.
Arch Pediatr Adolsce Med 2001; 155: 354-9.
Thompson N, Montgomery SM, Pounder RE, Wakefield AJ. Is measles vaccination
a risk factor for inflammatory bowel disease. Lancet 1995; 345: 1071-4.
Lavy A, Brodie E, Reif S et al.. Measles is more prevalent in Crohn's
disease patients. A multicentre Israeli study.Dig Liver Dis 2001; 33: 472-6.
Halsey NA, Hyman SL. Measles-mumps-rubella vaccine and autistic spectrum
disorder. Pediatrics 2001; 107: 1-23.
Wakefield AJ, Montgomery SM, Pounder RE. Crohn's disease: the case for
measles virus. Ital J Gastroenterol 1999; 31: 247-54.
Wakefield AJ, Montgomery SM. Measles virus and inflammatory bowel disease:
an unusually tolerant approach. Am J Gastroenterol 2000; 95: 3-5.
Montgomery SM, Morris DL, Pounder RE. et al.. Paramyxovirus infections in
childhood and subsequent inflammatory bowel disease.Gastroenterol 1999;
Ekbom A, Adami HO, Helmick CG, Jonzon A., Zack MM. Perinatal risk factors
for inflammatory bowel disease: a case-control study. Am J Epidemiol 1990;
Ekbom A, Wakefield AJ, Zack M, Adami HO. Perinatal measles infection and
subsequent Crohn's disease. Lancet 1994; 344: 508-10.
Pardi DS, Tremaine WJ, Sandborn WJ et al.. Early measles virus infection is
associated with the development of inflammatory bowel disease. Am J
Gastroenterol 2000; 95: 1480-5.
Montgomery SM, Bjornsson S, Johannsson JH, et al.. Concurrent measles and
mumps epidemics in Iceland are a risk factor for later inflammatory bowel
disease. Gut, 1998; 42: A41.
Montgomery SM, Pounder RE, Wakefield AJ. Age of older siblings and
inflammatory bowel disease. Gut 1999; 44: A29.
Montgomery SM, Twamley SI, Morris DL, et al.. Birth order influences IBD
risk and phenotype. Gut 1999; 44: A29.