MMR and autism

1
MMR and autism

• Dr Mary Ramsay
• HPA Communicable Disease Surveillance Centre

2
MMR and autism

• Theories originate from group studying
  inflammatory bowel disease at RFHMS
• Link with measles infection (1993-4)
• Link with measles vaccine (1995)
• Same group then described children with bowel
  disease and autism (1998)
• Ten years of research on measles and IBD and five
  years of research on MMR and autism
• Vast majority of research rejects any link
• So, why are we still plagued by this proposal?

3
Evidence on measles infection and Crohns disease

• Crohns disease (not ulcerative colitis) more
  common in those born during a measles epidemic in
  Sweden
• Measles virus found in gut tissue of cases of
  Crohns (not ulcerative colitis)
• Hypothesis 1 persistent measles infection
  causes Crohns disease
• Not confirmed by Wakefields own study in England
• Lab techniques for detecting measles criticised
• Lab findings not confirmed in UK, USA and Japan
  (including one study by Wakefield)
• no evidence of supporting antibody responses in
  blood

4
Evidence on measles vaccine and Crohns disease
AND ulcerative colitis

• A higher rate of IBD found in a cohort of
  vaccinated people compared to an unvaccinated
  cohort
• Hypothesis 2 measles (not MMR) vaccine causes
  IBD
• Groups were not comparable / asked different
  questions
• Findings not supported by
• large European case-control study
• UK case-control study
• UK cohort study (conducted by Wakefield - see
  next hypothesis)
• Incidence of Crohns is not related to use of
  vaccine
• in UK and in Finland

5
Evidence on mumps infection andinflammatory
bowel disease

• Follow up of cohort of children born in 1970
• Association between IBD and wild mumps infection
  before 2 years of age
• Association between IBD and wild measles and
  mumps infection in the same year
• Hypothesis 3 IBD is caused by early mumps
  infection (? some interaction with measles
  infection)
• Unexpected finding based on parental recall of
  childhood illnesses at age 10 years (most mumps
  is asymptomatic)
• No evidence of mumps virus or high antibody
  prevalence in IBD cases
• No evidence of increased risk of IBD after MMR

6
Evidence on MMR vaccine and autism

• Case series of 12 children with non-specific
  bowel symptoms
• 10 with autistic-like features
• 8 of the 12 children reported a temporal
  association between onset of symptoms and MMR
  immunisation
• We did not prove a causal association between
  MMR vaccine and the syndrome
• Hypothesis 4 MMR vaccine causes novel gut
  disorder which leads to autism

7
Main criteria for establishing causality for a
vaccine adverse event

• Biologically plausible?
• Laboratory evidence of vaccine involvement?
• Coherence with other knowledge?
• Correlation between rates of disease and
  vaccination
• Increased risk after vaccination?
• Clustering in a post vaccination period
• Higher risk in vaccinated compared to
  unvaccinated
• Specificity of association?
• Particular syndrome in vaccine-linked cases
• Consistency across studies?

8
Biological plausibility

• Virus damages the gut, leakage of proteins and
  peptides accesses the brain and causes damage
• Gut leakage would be in both directions no
  evidence of protein losing enteropathy in
  children
• Liver removes peptides from blood stream
• Peptides cannot cross the blood-brain barrier
• Gershon M. Autism and the measles-mumps-rubella
  (MMR) vaccine. Revisiones Vacunas 2002 2 156-7.

9
Laboratory evidence of vaccine involvement

• Molecular studies detected fragments of measles
  virus by PCR using novel primers
• 75/91 patients with ileal lymphonodular
  hyperplasia
• 5/70 developmentally normal control patients
• Uhlmann V et al. J Clin Pathol Mol Pathol 2002
  55 0-6
• Little information on cases and controls
  including age and vaccination history. Some
  cases had single measles vaccine. Controls were
  mainly appendicectomies.
• No information on validation of virological
  methods
• PCR method prone to contamination but no
  information on how samples handled or stored

10
Laboratory evidence of vaccine involvement

• Abstract claiming that measles RNA fragments
  found were vaccine derived
• Sheils O et al. Royal College of Pathologists
  2002.
• Based on identification of single nucleotide
  polymorphism (guanidine at 7901) rather than
  sequencing of whole PCR product (standard method)
• This nucleotide is also present in several wild
  measles strains (www.ncbi.nlm.nih.gov/nucleotide)
• Wakefield forced to admit conclusions were
  incorrect at US congressional hearing

11
Laboratory evidence of vaccine involvement

• Legal process commenced using handful of test
  cases chosen by the claimants
• CSF was obtained from 6 of the lead cases, three
  tested positive for measles in Irish laboratory
• Controls from around 20 leukaemic patients, only
  one positive
• Samples tested by defendants experts - all
  proved negative
• Cases refused legal aid on ground of poor evidence

12
Laboratory evidence of vaccine involvement

• Identification of high levels of MMR antibody
  in children with autism. Immunoblotting detected
  high levels of antibody to the H antigen of
  measles virus. Singh et al. J Biomed Sci 2002
  9 359-364.
• Cases and controls not adequately described (some
  controls were adults)
• Immunological methods not validated and findings
  are implausible
• Why no response to other measles antigens (M and
  N)?
• Why no response to mumps and rubella antigens
  despite vaccination?
• Molecular weight of protein detected more
  consistent with that of human albumin

13
Coherence with other knowledge

• Is there a correlation between rates of disease
  and vaccination in many countries?
• Apparent rise in autism in UK and California
  coincided with the introduction of MMR
  (Wakefield, Lancet 1999 354949-950)

14
Coherence with other knowledge

• Gillberg C et al. Autism 1998 2 423-4. No
  increase in autism in cohorts with increasing MMR
  coverage in Sweden.
• Taylor et al. Lancet 1999 3532026-29. Rise in
  diagnosed autism in North Thames pre-dated MMR
  vaccine, continued while uptake remained
  constant.
• Kaye JA,et al. BMJ 2001322460-3. GP data showed
  no correlation between the uptake of MMR
  vaccination and the rapid increase in the risk of
  autism.
• Dales L et al. JAMA, 2001285 1183-5. Time
  trends in autism and MMR immunisation coverage in
  California and showed no correlation.

15
Increased risk after vaccination

• Is there clustering in a post vaccination period?
• Parental testimonies of onset of regressive
  autism in previously normal child shortly after
  MMR

16
Studies looking for clustering of cases after MMR

• Taylor et al. Lancet 1999 3532026-29. No
  clustering of onset after MMR. No difference in
  MMR vaccine uptake between cases and rest of UK
  population.
• DeWilde et al. Br J General Practice, 2001 51
  226-227. No difference in pattern of GP
  consultation for UK autistic children in 6 months
  after MMR. Therefore no evidence of onset of
  behavioural disturbance associated with MMR in
  autistic children.
• Makela A et al. Pediatrics 2002 110 957-963. No
  clustering of hospitalizations for autism after
  vaccination in Denmark

17
North Thames studyTaylor B, Miller E, Farrington
CP et al submitted 1999

• Relative incidence of autistic regression
• within 2 months MMR 0.92 (0.38-2.21)
• within 4 months MMR 1.00 (0.52-1.95)
• within 6 months MMR 0.85 (0.45-1.60)
• within 2 months MCV 1.24 (0.61-2.56)
• within 4 months MCV 1.31 (0.73-2.33)
• within 6 months MCV 0.99 (0.56-1.75)

18
Increased risk after vaccination

• Is there a higher risk in vaccinated children
  compared to unvaccinated?
• Geier M, Geier D. International Pediatrics 2003,
  Vol 18 108-113. Compared the number of passive
  reports of autism reported after MMR compared to
  the number of reports of autism reported after
  DTP. Estimated a high RR after MMR.
• Based on passive reports to VAERs. Not designed
  to test hypotheses, subject to bias from adverse
  publicity.
• Controls received DTP therefore different age
  to those receiving MMR

19
Studies looking for higher risk in vaccinated
children

• Farrington et al. Vaccine 200119(27)3632-5.
  Reanalysis of North Thames data - no association
  between autism and MMR at any time interval.
• Madsen KM et al. NEJM 2002 347 1477-82. Risk of
  autism same in unvaccinated and vaccinated
  children in Danish population based study.
• Peltola et al. Lancet 1998 351 1327-28. No
  evidence for MMR vaccine being associated with
  inflammatory bowel disease or autism (14 year
  prospective study)

20
Wakefields response to rejection of hypothesis

• MMR vaccine might cause autism, but that the
  induction interval need not be short
• Therefore no clustering of cases after receipt of
  MMR
• MMR causes autism but only if co-factors
  present
• Therefore, no correspondence necessary between
  MMR uptake and autism incidence
• Main evidence is now the alleged novel syndrome
  of autistic entercolitis presenting with
  regression
• Presents at any stage after vaccination

(antibiotics, intercurrent infection, atopy,
family history of autoimmunity, etc)
21
Specificity of the association

• Is there a new syndrome autistic enterocolitis
  with onset of regressive symptoms after MMR?
• Fombonne E. Lancet 1998 352 955. No evidence of
  an association between IBD and autism.
• Fombonne E. Pediatrics 2001 108 e58. No
  evidence for a new variant of autism in recent
  years. No change in age at parent concern or age
  at regression for cases who had received MMR.
• Taylor B et al. BMJ 2002 324 393-6. No evidence
  of increased rate of regression or bowel symptoms
  in children who developed autism after MMR.

22
Percentage of Cases with Bowel Symptoms
by MMR Status
P-value 0.36
25.0
20.0
15.0
10.0
5.0
0.0
MMR Post - Parental
MMR Pre-Parental
No MMR
Concern
Concern
23
Percentage of Cases with Regression
by MMR Status
P-value 0.83
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
No MMR
MMR Pre-Parental
MMR Post - Parental
Concern
Concern
24
Main criteria for establishing causality for a
vaccine adverse event

• Biologically plausible? NO
• Laboratory evidence? NO
• Coherence with other knowledge? NO
• Increased risk after vaccination? NO
• Specificity of association? NO
• Consistency across studies? NO

25
Summary of evidence on MMR and autism

• Evidence to support the hypothesis that MMR
  causes autism is weak and inconsistent
• Hypothesis falls down on all major criteria
• Strong evidence to reject hypothesis now
  published
• However
• MMR coverage has fallen
• Major demand for single vaccines
• Control of measles is under threat

26
Demand for single vaccinesScientific basis

• Originates from Wakefields comments at press
  conference in 1998
• Suggested giving single vaccines one year apart
• Probably based on one study showing an
  association between inflammatory bowel disease
  and having natural mumps and measles in the same
  year (Montgomery SM. Gastroenterology 1999, 116
  796-803.)

27
Demand for single vaccinesDismissal of
scientific basis

• First RFH study showed an association between IBD
  and SINGLE measles vaccine (Thompson NP et al.
  Lancet 1995 345 1071-4.)
• Another RFH study showed an association with
  mumps infection below the age of two years
  (Montgomery SM. Gastroenterology 1999, 116
  796-803.)
• Therefore giving single vaccines one year apart,
  with measles vaccine first
• Still exposes child to vaccine implicated in IBD
• mumps protection could be delayed by 1-5 years
• increase the risk of wild mumps at an early age

28
Demand for single vaccinesPolitical basis

• Impossible to show lack of an extremely rare
  association
• Advise single vaccines just in case
• Current recommendations are not working
• Advise single vaccines to avoid imminent
  epidemics
• Parents should be given choice
• Allow single vaccines for informed parents

29
Advise single vaccines just in case Dismissal
of Political basis

• This is not a case of lack of evidence
• This is strong evidence of no association
• Why should single vaccines be intrinsically
  safer?
• Data on population use is limited
• What if MMR prevented autism?
• MMR is the safer option
• Case control study of autism from Japan showed RR
  of 5.33 for single measles vaccine (Takahashi H
  et al. Jpn J Infect Dis 2003 56 114-7)

30
Advise single vaccines to avoid epidemics
Dismissal of Political basis

• Would this lead to improved coverage?
• Current coverage
• at two years is 79 (fall of 13)
• at five years of age is 90 (fall of 4)
• If all those refusing MMR had single vaccines -
  potential gain of between 4 and 13
• What would be the drop-out rate for all three
  vaccines (given one year apart)?
• similar to that for DT4 pre-school? 21

31
Allow single vaccines for informed parents
Dismissal of Political basis

• Single vaccine choice is actually misinformed
• More data on safety of MMR than single vaccines
• Parents have been misled by balance of media
  reporting (Report from Cardiff School of
  Journalism, Media and Cultural Studies)
• Allowing opt-out would provide mixed messages
• Particularly to the majority of parents who have
  accepted MMR
• Mixed schedules could be damaging to both
  individual and population
• Experience in Greece with rubella
  (Panagiotopoulos T, BMJ 1999 319 1462 1467)

32
Vaccination strategies to control congenital
rubella syndrome (CRS)

• Selective vaccination of girls
• allows acquisition of natural immunity in
  childhood
• direct protection of women of childbearing age
• Universal vaccination
• aim to eliminate rubella infection
• indirect protection of women of childbearing age
• Greece
• MMR used in private practice
• rubella used in public sector (along with measles
  and mumps single vaccines) with poor control

33
Effects of routine infant vaccination with
rubella vaccine

• Reduction in number of cases
• reduced risk of infection
• increasing susceptibility in older age groups
• increased age at infection
• Potential for
• increase in cases in adult women
• ? increase in cases of CRS

34
Predicted incidence of CRS Selective and
universal vaccination programmes
following Anderson Grenfell, 1986
35
Congenital rubella cases and rubella
notifications Greece 1991-1998 (source WHO)
36
Problems with recommending single vaccines

• Giving the vaccines separately will be harmful as
  children and their contacts would be exposed for
  longer
• No evidence to support appropriate intervals
• Drop-out rates and partial schedules will
  increase
• Single antigen vaccines are likely to be in short
  supply
• No other country using all three vaccines
• Companies unlikely to maintain license and
  control
• Enforced incomplete courses (cf. Td, D and T)
• Consistent advice and recommendations are needed
  (cf. pertussis experience)

37
UK experience with pertussis

• Concern about safety of pertussis component of
  vaccine
• Supported by some health professionals
• Parents given choice to opt-out
• Collapse in vaccine coverage for pertussis
• Led to major epidemics
• Knock-on effect on other vaccines
• Major factor associated with non-vaccination was
  inconsistent advice from health professionals

38
Whooping cough cases and vaccine coverageEngland
and Wales 1940-2002
Immunisation introduced
80
40
0
39
Immunisation coverage England and Wales,
1966-2002
Vaccination coverage
40
Conclusion

• MMR and autism hypothesis has been firmly
  rejected
• Evidence is in favour of continuing MMR without
  single vaccine option
• Health professionals need to be consistent
• Our challenge is to redress the media balance