The Schafer Autism Report
The World's Widest Read Autism Publication On or Off the Internet
Healing Autism:
No Finer a cause on the Planet
| The Schafer Autism Report The World's Widest Read Autism Publication On or Off the Internet | Healing Autism: |
MMR and Acquired Autism
(Autistic Enterocolitis)
- A Briefing Note
| (vaccine) | Reported adverse events | Reported serious adverse events | Reported deaths | % of total events reported as serious** | % of adverse events reported as deaths** |
| Dipther Tet | 1,492 | 189 | 15 | | |
| DTAP | 10,348 | 1,422 | 283 | | |
| DipTetPert | 21,163 | 3,286 | 794 | | |
| DTPH | 6,212 | 928 | 254 | | |
| Flu | 15,351 | 2,082 | 324 | | |
| Hepatitus B | 32,209 | 4,676 | 662 | | |
| HibV | 21,726 | 3,905 | 932 | | |
| Measles | 414 | 61 | 7 | 15% | 2% |
| Measles M | 34 | 25 | 2 | 74% | 6% |
| MMR | 20,974 | 2,586 | 132 | 12% | 1% |
| Measles R | 117 | 23 | 0 | 20% | 0% |
| Mumps | 54 | 19 | 3 | 35% | 6% |
| Polio live or | 24,702 | 3,541 | 970 | | |
| Pneumococ | 5,841 | 712 | 95 | | |
| Rubella | 685 | 100 | 1 | 15% | 0% |
| Tetanus Dip | 9,566 | 520 | 12 | | |
| Varicella | 12,635 | 590 | 31 | | |
| TOTALS* | 201,815 | 27,768 | 4,965 | 14% | 2% |
Notes: * totals include a number of other vaccines, not included in the table,
(UK Department of Health's "Top 10 Truths")
| (Department of Health "Truth") | (Critical Response of Parents) |
| MMR is safest way to protect children | Does not address the alleged damage |
| Over 500m doses of MMR have been used in over 90 countries | Almost all those countries have no autism database. Only US has good data - and this shows a steep rise in autism |
| No country in the world recommends single vaccines | No country in the world has yet acknowledged that there may be an MMR/autism link, either, but that may yet follow in time. Some countries permit single vaccines as a choice. |
| Children who are not immunised with MMR increase the chance of infection in others. | True. But those children could still receive single vaccines. And there may yet be a massive loss of confidence in all vaccination, if the children win in the High Court. It would therefore be prudent to think of this possibility, and permit choice now. |
| The evidence is that MMR does not cause autism or IBD (a number of studies are quoted, but only those which suit the Department's stance) | There is evidence that suggests that it may do. Every one of the quoted studies that "disproves" an MMR/autism link can be flawed (see elsewhere in this document). |
| Wakefield et al in 1998 said "We did not prove an association". | True. The research is still unfolding. Time did not stop in 1998. |
| Single vaccines put children at risk | The Department's argument is based upon a supposition that some children would not complete the full course of vaccines. But if the children win in the High Court, and the Department is shown to have misled the public (either unknowingly or knowingly), the damage will be far greater. And already, some children are avoiding any measles vaccine. The Department's argument is already having a perverse consequence, and may eventually massively backfire.. |
| MMR was thoroughly tested before introduction into the UK in 1988. | In the context of adverse outcomes with an insidious long-term onset, MMR was not properly tested. Advice at the time to explore possible adverse effects was not followed up. By disputing historical facts, the Department reveals its bias. |
| Two doses of MMR are needed to protect children. | The efficacy of MMR in terms of preventing measles is not the point at issue. |
| There are very few children with genuine contraindications. | This does not address the MMR/autism link. It also does not square with the manufacturer's own information sheets, which imply a substantial number of possible adverse effects. |
The Department of Health's "Top 10 Truths" leaflet ends with the reassuring statement, "All of the above are correct"! The above critique suggests that the "truth" is nowhere near clear-cut, and the Department's position is thus exposed as artificial and one-sided.
(UK Department of Health's "Top 10 Myths")
| (Department of Health "Myth") | (Critical Response of Parents) |
| Getting protection by catching the disease is better. | This is not the issue in dispute. |
| Three viruses given at the same time is too much for children. | It may yet prove to be. The Department has no evidence (in the context of the MMR/autism debate) to the contrary, in relation to live viruses. |
| Other countries recommend that MMR is given as separate vaccines. | Of course they don't. Perhaps this is because no country has yet woken up to the problem. As yet, there is insufficient evidence to alter this position. |
| Measles, mumps and rubella are rare in the UK so there is no need to immunise. | This is not the issue in dispute. |
| MMR causes autism and bowel disease. | There is evidence pointing towards an MMR/autism/IBD connection. Until this area is thoroughly researched, it is scientifically untenable to rule it out. |
| There was a scientific paper that linked MMR and autism/IBD | There have now been a number of such papers. They form part of an unfolding story. |
| Giving MMR as separate vaccines reduces the risk of side effects. | It is not possible to prove/disprove this until proper clinical research has been funded and conducted. |
| The vaccine was not properly tested. | In the context of the MMR/autism debate, and the alleged link, this is factually true, and it is extraordinary for the Department to claim otherwise. Even the Department cannot re-write history. |
| My child has already received one dose, so does not need a second dose. | This is not the issue in dispute. |
| My son does not need protection against rubella, my daughter does not need protection against mumps. | This is not the issue in dispute. |
The Department of Health's leaflet ends, "All of the above are wrong". In the view of the parents, of the "Top 10 Myths", four are irrelevant to the debate about an MMR/autism link, one statement about a "Myth" is factually incorrect, and the remainder can readily be disputed because the research has not been completed, or in some cases even commissioned, to decide the issue either way.
The position in the US is no different. In summer 2002, the US Center for Disease Control (CDC) updated its "Frequently Asked Questions" (FAQs) on the MMR/autism issue. It asked the question: "What have studies found regarding MMR vaccine and autism?".
Its answer was "Epidemiologic studies have shown no relationship between MMR vaccination in children and development of autism". However, what it did not acknowledge, or discuss, was that "studies" in the original question should have included both clinical and epidemiological studies, with greatest weight being attached to clinical findings. Its answer ducked the issue of clinical studies, focussing solely on epidemiological studies (see later for a critical review of these).
10. Position of US Center for Disease Control on MMR/Autism
The position of the US Center for Disease Control is summarised as follows (taken from their website in February 2002, but believed to be unchanged as at February 2003):
Ÿ Is there any scientific evidence that provides a link between autism and vaccination? - To date there is no convincing evidence that any vaccine can cause autism or any kind of behavioural disorder. A suspected link between MMR vaccine and autism has been suggested (but this).......may simply be an.....unrelated chance occurrence.
Ÿ Is there a theoretical possibility that there is a connection between autism and MMR vaccine, or any other vaccine? - If measles vaccine or any other vaccine causes autism, then it would have to be a very rare occurrence, since millions of children have received vaccines without ill effects.
Ÿ What are the known side-effects associated with MMR? - About 5-15% of vaccinees may develop a fever 5-12 days after MMR, and 5% may develop a rash (comment - not clear if this means 5% within the 15% or 5% plus the 15%). Central nervous system conditions, including encephalitis and encephalopathy, have been reported with a frequency of less than one per million doses administered
Ÿ What is the federal government doing to protect the health of persons who receive MMR? - There are no proven data to suggest that measles vaccine will increase the risk of developing autism or other behavioural disorders.
Comment: the above is neither comprehensive nor balanced, and its one-sided reassurance is therefore unhelpful. The details of the above could even be challenged on the grounds of factual accuracy. Point one is particularly threadbare.
11: The Parents Have Seen What They've Seen.......
It is not in dispute that vaccines have saved millions of lives. The MMR/autism parents are not anti-vaccination in principle. These parents all took children to be vaccinated. We all recognise the need to protect children from diseases.
But saving lives from diseases doesn't justify ruining significant numbers of lives from unrecognised and unmonitored vaccine damage.
It is also felt by many parents that the mantra "the benefits of vaccination outweigh the risks" has become increasingly skewed by
Ÿ (a) occasionally overstating the dangers of diseases, citing experience of diseases from poor and underdeveloped countries, or UK experiences from half a century ago, or pointing to recent deaths (e.g. Ireland) where other factors played a major part, or
Ÿ (b) grossly underplaying or dismissing outright any risks from vaccination. This latter has been aided by the extremely poor monitoring of adverse outcomes, and by the authorities strenuously refusing to accept that an adverse outcome was the result of a vaccine.
All affected parents are in the privileged position of having watched their child degenerate. It is a powerful first-hand experience. Comparing notes results in finding that other parents have undergone extremely similar experiences. Unfortunately, such experiences are not part of a scientifically-controlled study, so are routinely dismissed by the Department of Health as anecdotal.
Ÿ Usually there appears to be a very gradual degeneration over many weeks and months, not an acute event, more akin to (eg) the onset of cancer than the rare acute reactions to vaccines seen in the past.
Ÿ But all the attention of the past upon possible adverse reactions to vaccines has focussed upon acute near-immediate events.
Ÿ The onset of gut/bowel problems and hyperactivity have accompanied the onset of autism. Some link between them is therefore likely, even without detailed research.
Ÿ An anecdote is an anecdote. A consistent pattern of anecdotes is much more powerful. What we have is a consistent detailed pattern of reports from parents. The importance of this pattern has been ignored by the Department of Health.
PART B
THE COSTS OF AUTISM
12: The Financial Costs - Autism Is Costing £$Billions
Quite apart from the immense social costs of autism, there are the huge financial costs. Autism effects every UK and US taxpayer. In the UK, the costs comprise:
Ÿ Health costs - specialist hospital visits, GP visits, prescriptions, exclusion diet costs
Ÿ Education costs - special schools, extra teachers, extra teaching assistants, extra training
Ÿ Transport costs - taxis plus drivers and escorts, plus local authority management costs, plus environmental/congestion costs of extra traffic
Ÿ Social Services costs - respite care costs, transport, management, inspection, reviews
Ÿ Loss of earnings of parents acting as carers
Ÿ Social Security costs - carers allowances, disability living allowances
Ÿ Inland Revenue costs - loss of earnings of parent, loss of revenues from child when he/she reaches earning age
Ÿ Wider economic costs - loss of gross domestic product to the national economy
It would be interesting to know if the UK Treasury had a view on these costs, and whether sufficient resources were being devoted to investigating acquired autism and other forms of autism, as they represent a significant loss to the wider national economy. Is autism too important to be left to the Department of Health?
13: Estimates
In June 2000 a study for the Mental Health Foundation found that
Ÿ the annual costs of autistic disorder in the UK were at least £1 billion
Ÿ individual lifetime costs per child affected could run to £2.94 million each.
The full costs, taking into account wider economic costs, are probably considerably higher still.
14. Failure To Monitor Increases In UK Autism Numbers
Ÿ There has been a consistent argument on the part of the authorities, and those seeking to defend MMR, that the apparent rise in autism may be largely a matter of better recognition. This has received some backing from autism researchers. But where hard UK or US data is available, increases are far too steep, and in far too short a timescale, to be credibly ascribed to better recognition alone..
Ÿ For this to be "better recognition" or "improved diagnosis", this would have required these children to have been missed, simultaneously, by their parents, their relatives, their doctors and their teachers in the past This is simply not credible. For example, the increase in autism 1992-99 in Wakefield, West Yorkshire, local education authority was from 5 cases to 111 cases. If increased autism is down to better recognition, it would mean that, back in 1992, there really were 111 cases, but only 5 were recognised, and the remaining 106 were missed, and by all the parties - parents, doctors, health visitors, teachers - concerned. This is completely implausible.
Ÿ Undoubtedly there has been some degree of better recognition and reclassification, following introduction of ICD-10 (international classification of diseases/disorders) criteria in 1992, and DSM-IV (diagnostic statistics manual) criteria in 1995. But this will account for only a minority of the growth.
Ÿ The UK DoH has failed to monitor autism, and is still failing to (despite a specific 1997 recommendation of the House of Commons Health Committee to do so). Is it now afraid of what it might find? If it does decide to monitor autism, will it find that numbers are high and then claim it has always been so?
Ÿ UK Health Boards/Authorities are also failing to monitor autism locally. Health Boards/Authorities have little data and no consistent approach. At the health authority level, official figures vary wildly, by factor of 300-fold, i.e. 300-times (not 300%). The data is an extraordinary mess.
Ÿ In the year 2000, only 1 in 6 UK Boards/Authorities had any credible figures at all. Most used estimates from textbooks.
Ÿ The Scottish schools census now includes autism. The census commenced in 1998. The 1999 census showed 18% increase over the 1998 census. The 2000 census showed a 31% increase over the 1999 census. The 2001 census will report during mid-2002.
Ÿ There are other indications of the level of increases: Kaye et al paper (see later) found a sevenfold increase 1988-99 in UK. An unpublished 1999 paper by Dr. Fiona Scott, Autism Research Unit, Cambridge, indicated autism at eleven times the expected level (1 in 174) - see later.
Ÿ The 2001 Medical Research Council review found autism to be at 1 in 166, many times higher than hitherto thought. Sixteen studies published between 1966 and 1991 found rates of between 1 in 3030 and 1 in 625. A rate of 1 in 166 is nearly four times higher than 1 in 625, itself the highest of these sixteen, and from a relatively-recent study in 1983.
The repeated official line that the apparent increase is down to better recognition may therefore be little more than a counsel of complacency.
In December 2002, a Parliamentary Written Question (84502) confirmed that there is now in place a "Good Practice Guidance on Autistic Spectrum Disorders", in the UK, published by the Government's Departments of Education & Skills and of Health. This is intended to raise awareness amongst schools and local education authorities. However, it is probably just one of many thousands of such well-intentioned documents, is non-statutory, and is probably lost in the stream of paper raining down on local government from central government.
UK schools and local education authorities have a duty to identify, assess and make suitable provision for children with special educational needs. However, there seems to be no duty upon either the health authorities at the local level or the Department of Health at Government level to improve the data position over autism - doubtless to the latter's relief. Perhaps centrally-collated figures showing steep increases would beg uncomfortable questions as to the causes. The UK Department of Health seems to regard autism as a problem for local education authorities - not for the Department.
15. "Now Almost Everyone Knows Someone Who's Autistic"
Autism was a very rare condition, but is now almost regarded as commonplace. Very many cases are now of late- onset autism, whereas almost all used to be cases from birth. We have to ask why this is.
Some UK research noted the sharp increases in autism in the 1990s. A paper by Powell et al, Department of Public Health and Epidemiology, University of Birmingham, UK, Changes in the Incidence of Childhood Autism and Other Autistic Spectrum Disorders in Pre-School Children from Two Areas of the West Midlands, UK, was published in Developments in Medicine and Child Neurology, September 2000. This looked at the incidence of childhood autism and ASD in pre-school children between 1991 and 1996.
The study found that there were year-on-year increases in classical autism during this period of 18%, but for "other ASDs" the annual increase was no less than 55%. But the study then concluded that this was due to clinicians being increasingly able or willing to make a diagnosis. The possibility of an underlying genuine increase, and any follow-on question as to causes, does not appear to have occurred to the study team.
But parents of children believe to have been damaged by MMR strongly believe that part of the increase is down to a new phenomena, autistic enterocolitis.
It is not the autism of the past. Such a severe acquired regressive syndrome after a normal early childhood would have been noticed in the past by parents, and recognised medically, and also reflected in much higher historic rates of prevalence/incidence.
In the parents' view, there is clear evidence of recent dramatic rates/increases:
Ÿ examples - an East Surrey 1/69 rate amongst three year old boys, 1/139 rate amongst three year old boysŸ Bromley Autistic Trust figures show a 1990-94 increase of 280% over 1980-84 figures (source: personal communication of 16/9/99 from Miss C. M. Povey, Services Director, Bromley Autistic Trust)
Ÿ Wakefield LEA autism pupils up from 5 to 111 in seven years (source: survey by David Brown, a specially-seconded headmaster from the Park School, Wakefield, on behalf of Wakefield Local Education Authority, 1999)
Ÿ Telford health data up from 4 new cases per year in 1990 to 17 per year 1998 and again 1999 (source: personal communication of 20/11/00 by Dr F. R. J. Hinde, Consultant Paediatrician, Princess Royal Hospital, Telford)
Ÿ As noted, Scottish schools census up 18% in one year (1999 vs. 1998), and then a further 31% in the next year (2000 vs. 1999); (source: Scottish Annual School Censuses, available from Scottish Education Office, tel 0131 556 8400)
On December 22nd 2002, the UK Observer newspaper carried a report on the apparent epidemic of behavioural problems amongst UK schoolchildren. Whilst not confined to autism (the report cited hyperactivity and attention-deficit disorder), the Observer's report suggested a steep rise in the incidence of problems. Figures obtained by the newspaper suggested that numbers of schoolchildren with attention-deficit disorder (ADD) or attention-deficit hyperactivity disorder (ADHD) had reached 345,000, and that one child in twenty between the ages of 6 and 16 years had one or other condition. The Observer also found out that prescriptions for Ritalin, to counter these disorders, had increased markedly, from 91,100 in 1997 to 208,500 in 2001.
In the US, the Brown University Child & Adolescent Behavioural Letter (18(3): 1: 304, 2002) carried the following details:
Ÿ A study into attention deficit hyperactivity disorder (ADHD) was undertaken, based on parent and teacher reports concerning 6,099 children in 17 public elementary schools. The study was undertaken by researchers working for the National Institute of Environmental Health Sciences in North Carolina
Ÿ When the researchers surveyed parents in a typical county of rural and suburban communities - Johnston County, North Carolina - the parents reported that more than 15% of boys in grades 1st through 5th had a diagnosis of ADHD, with about 10& (i.e. two-thirds of those diagnosed) receiving medication.
Although ADHD is not autism, it may share some common causal pathways, particularly multiple food allergies and gut permeability. The finding is thus of interest to the MMR/autism debate.
16. University of Cambridge Research
On 18/2/01, the UK Sunday Telegraph reported on research undertaken by Dr. Fiona Scott at the Autism Research Centre at the UK University of Cambridge. The research, Prevalence of Autism Spectrum Conditions in Children Aged 5-11 Years in Cambridgeshire UK, by Scott, Baron-Cohen et al, which is due to be published shortly, was undertaken across schools in Cambridgeshire.
The study aimed to establish prevalence of the broader autistic spectrum, including Asperger syndrome in 5-11 year olds in Cambridgeshire, UK. Cases of diagnosed autism spectrum condition in children who were in Cambridgeshire schools and aged 5-11 on 31st December 1999 were sought out using public records, screening instruments, educational psychology and special educational needs coordinator records.
It found that:
Ÿ One in 175 (58/10,000) children was autistic, whereas previous studies had pointed to a rate of 1 in 2000 (5/10,000)
Ÿ This was 11 times higher than the rate of classic autism, but in line with other recent national and international rates for the broader spectrum.
Ÿ In responding mainstream schools, the prevalence was 1 in 300. In the responding special schools, the prevalence was 1 in 8.
Ÿ Extrapolated across the UK, that would imply 30,000 primary school (age 5-11) children with autism
Ÿ The overall sex ratio of the children was 4 to 1 male to female, but in mainstream schools it as 8 to 1.
Ÿ Linking these rates to estimated costs of education and care for sufferers would give a figure of as high as £5 billion per year, year after year. The Cambridge autism figures were described as "if anything an under- estimate". They included only children with a definite clinical diagnosis. Any child who had only been "statemented" (= educational needs-assessed) as autistic, but not yet clinically diagnosed, was not counted
Ÿ One in eight children with special educational needs was suffering from some form of autistic spectrum disorder. The increase of actual numbers over previously-assumed numbers would have enormous cost implications for central and local Government
Ÿ A year-2000 report for the UK Mental Health Foundation by Professor Martin Knapp for the UK Institute of Psychiatry used the earlier "textbook" rate of autism of 5/10,000 to put the total UK economic cost of autism at £1bn. The Knapp report estimated the lifetime cost of a severely-affected child at £3m, for a high-functioning autism child at £0.8m, and for an Asperger's syndrome child at £0.5m. The revised £5bn per year estimate is based upon these costs.
17. University of Sunderland Research
An unpublished study by the UK University of Sunderland found a tenfold increase in diagnosis of autism, during the years 1989-93.
18. UK National Autistic Society Estimates
The NAS issued a factsheet in early 1997 which gave the following prevalence rates:
Ÿ People with Kanner syndrome (IQ less than 70) 5/10,000, or 1 in 2,000
Ÿ Other spectrum disorders (IQ less than 70) 15/10,000, or 1 in 666
Ÿ Asperger's (IQ 70 or above) 36/10,000, or 1 in 278
Ÿ Other spectrum disorders (IQ 70 or above) 35/10,000, or 1 in 286
Combined total of above four groups 91/10,000, or 1 in 110
The above implies a very high level of autism in the UK, and the previously-described studies seem to bear this out.
The NAS reach its 91 in 10,000 or 1 in 110 rate by taking the Wing & Gould study (Camberwell, London) of 1979, which looked at children with an IQ of under 70 and found a rate of 20 per 10,000, and adding this to the study by Ehlers & Gillberg (Sweden) of 1993 which looked at autistic children with an IQ of over 70 and found a rate of 71 per 10,000 (1 in 141).
The 91/10,000 rate is thus "merged data", collected in two different countries and some years apart, and thus needs to be treated with caution, particularly if rates have since been rising further. The Wing & Gould study is now over two decades out of date, and also pre-dates MMR introduction into the UK.
19. Report by Fiona Loynes, UK All-Party Parliamentary Group on Autism, Dec. 2001
The purposes of this report included:
Ÿ To establish numbers of children with autistic spectrum disorders
Ÿ To learn whether UK local education authorities believed there had been a recent increase in the last five years
Ÿ To ascertain whether LEAs routinely collected data
The findings included the following:
Ÿ 100 out of 115 LEAs reported an increase in autism in the past five years. Some reported small increases, others reported far higher increases, in one case by 77%.
Ÿ The study compared the expected prevalence rate of all autistic spectrum disorders in each LEA (91 in 10,000 or 1 in 110) with the actual recorded number of children with ASD and a Statement of Educational Needs (21 in 10,000 or 1 in 476). If the estimated numbers are correct, then the implication is that 75% of children with autism do not become included in the Statement data, because they have no Statement.
Ÿ Only 44 out of the 100 LEAs reporting an increase had actual data. Some of these reported dramatic increases, up to 400% in four years.
20. Report, "Autism In Schools - Crisis or Challenge", National Autistic Society UK, May 2002
Ÿ This report was complied from the findings of a survey carried out in seven local education authorities across England, Wales and Scotland, although the Scottish findings were reported separately. The England and Wales survey involved 373 individual surveys, with a response rate of over 30%, covering a pupil population of 133,000. The study found that:
Ÿ 1 in 86 children in mainstream schools had special educational needs that were related to ASD.
Ÿ The rate of ASD is three times higher in primary than in secondary schools. In primary it is 1 in 80, in secondary it is 1 in 268.
Ÿ This is in addition to children with ASD in special schools. In special schools, 1 in 3 children has ASD-related needs.
21. Is Autism Increasing? - Some Recent Official UK Pronouncements
These are some recent, and sometimes self-contradicting, statements:
Ÿ "There is no good evidence that the frequency of autism has increased since the introduction of MMR" - Tessa Jowell, then Minister for Public Health, October 1997 (personal communication to David Thrower)
Ÿ "The true incidence of autism is uncertain" - Sir Kenneth Calman, then Chief Medical Officer, March 1998
Ÿ The apparent rise in autism in the UK began more than ten years before the introduction of MMR" - Tessa Jowell, in June 1998
Ÿ "Rates of autism are rising, but not because of MMR" (Committee on Safety of Medicines, June 1999)
Ÿ "There is no robust data on the prevalence of autism before and after MMR's introduction" - Brent Taylor, in a June 1999 study heavily quoted by the Department of Health
Ÿ "Numbers of cases of autism are rising, but the reason for this is unclear" - John Hutton, Minister for Public Health, December 2000
Ÿ "Methodological differences between studies, changes in diagnostic practice and public and professional awareness are likely causes of increases in prevalence. Whether these factors are sufficient to account for increased numbers of identified individuals, or whether there has been a rise in actual numbers, is as yet unclear" - Medical Research Council 2001 review, quoted by the Scottish Parliament Expert Group May 2002.
Ÿ "Two thirds of (surveyed) teachers felt that there were more children with ASD now than five years ago. This (is) consistent across age groups and in all types of education provision, special and mainstream" (Report of the National Autistic Society, May 2002)
22. Autism In The USA
Ÿ The UK Department of Health is fond of saying how MMR is safely used in 32 countries, including the USA, as though its use elsewhere is proof in itself that it is safe. Recent claims have even referred to 100 countries. But the USA, at least, has clear evidence of an autism epidemic. Other countries may also be becoming aware of increases, for example Finland, where a 400% increase in cases has been alleged since was MMR introduced.
Ÿ The US has IDEA (Individuals with Disabilities Education Act). This picks up numbers of schoolchildren with developmental problems. Autistic pupils are up from 12,222 to 78,717 between 1992-1993 and 2000-2001 (Source: US IDEA State data).
Ÿ To the above total also has to be added a further 15,581 cases of autism amongst children aged 3-5 years, as at year 2000 (this number will have since increased further).
Ÿ There have been huge increases in some States between 1992-1993 and 1999-2000 - up 885% in Alabama, 529% in Connecticut, 435% in Florida, 513% in Idaho, 636% in Kansas, 561% in Minnesota, all in just seven years (Source: US State data, Individuals with Disabilities Education Act)
Ÿ It is also interesting that individual towns such as Round Rock, Texas, are reported to be up from 6 cases to 115 cases in eight years - very much like Wakefield Local Education Authority in the UK (up from 5 to 111 in seven years). On the face of it, this suggests that UK increases may very closely match those in the USA.
Ÿ It has been alleged that Brick Township (New Jersey) has manifested an "autism cluster". Some 40 of Brick Township's 6,000 3-10 year olds have autistic spectrum disorder. It has made Brick Township the "autism capital of the USA" (but note, East Surrey rates in the UK are higher still). In Brick Township, Federal investigators collected data on surface and ground water, sites of industrial spillages and waste dumping, and also ensured that there had been correct diagnosis of the actual children. They have found nothing untoward. Their findings were reported in April 2000.
Ÿ The following is taken from the statistics produced by the Department of Education in the United States, for numbers of children aged 6-21 served by IDEA (Individuals With Disabilities Discrimination Act) who have autism. It compares the increase over the eight years between 1992-93 and 2000-01:
| State | 1992-1993 | 2000-2001 | Percentage Increase |
| Alabama | 68 | 765 | 1,025 |
| Alaska | 8 | 195 | (almost infinite) |
| Arizona | 199 | 1,119 | 462 |
| Arkansas | 30 | 671 | 2,137 |
| California | 1,605 | 10,557 | 558 |
| Colorado | 14 | 453 | (almost infinite) |
| Connecticut | 164 | 1,225 | 647 |
| Delaware | 15 | 263 | 1,653 |
| District of Columbia | 0 | 103 | (infinite) |
| Florida | 582 | 3,926 | 575 |
| Georgia | 262 | 1,916 | 631 |
| Hawaii | 52 | 276 | 431 |
| Idaho | 39 | 291 | 646 |
| Illinois | 5 | 3,103 | (almost infinite) |
| Indiana | 273 | 2,621 | 860 |
| Iowa | 67 | 537 | 701 |
| Kansas | 74 | 619 | 736 |
| Kentucky | 38 | 864 | 2,174 |
| Louisiana | 409 | 1,145 | 180 |
| Maine | 37 | 444 | 1,100 |
| Maryland | 28 | 1,933 | (almost infinite) |
| Massachusetts | 493 | 575 | 17 |
| Michigan | 288 | 4,075 | 1,315 |
| Minnesota | 296 | 2,448 | 727 |
| Mississippi | 0 | 385 | (infinite) |
| Missouri | 336 | 1,589 | 373 |
| Montana | 20 | 163 | 715 |
| Nebraska | 4 | 337 | (almost infinite) |
| Nevada | 5 | 394 | (almost infinite) |
| New Hampshire | 0 | 342 | (infinite) |
| New Jersey | 446 | 2,925 | 559 |
| New Mexico | 16 | 225 | 1,306 |
| New York | 1,648 | 5,943 | 260 |
| North Carolina | 786 | 2,374 | 202 |
| North Dakota | 9 | 118 | (almost infinite) |
| Ohio | 22 | 2,217 | (almost infinite) |
| Oklahoma | 31 | 666 | 2,048 |
| Oregon | 37 | 2,516 | 2,516 |
| Pennsylvania | 346 | 3,304 | 855 |
| Puerto Rico | 266 | 473 | 78 |
| Rhode Island | 19 | 309 | 1,526 |
| South Carolina | 141 | 852 | 504 |
| South Dakota | 36 | 227 | 531 |
| Tennessee | 304 | 935 | 208 |
| Texas | 1,444 | 6,023 | 317 |
| Utah | 105 | 584 | 456 |
| Vermont | 6 | 160 | (almost infinite) |
| Virginia | 539 | 1,983 | 268 |
| Washington | 476 | 1,620 | 240 |
| West Virginia | 101 | 312 | 209 |
| Wisconsin | 18 | 1,823 | (almost infinite) |
| Wyoming | 15 | 94 | 527 |
| Total | 12,222 | 78,717 | overall increase 644 |
(Source: Individuals With Disabilities Education Act data, US Department of Education. Note: Where increases are from a very low base figure, these have been expressed as "almost infinite".)
Ÿ For every two cases there were in 1993, there are now thirteen. And the latest 2000-2001 figures represent a single-year increase of 20% over 1999- 2000
Ÿ The current estimate for the year-end of 2002 is 94,000-95,000.
Ÿ It seems obvious that the US has an autism epidemic. The UK is a similar health environment to the US, so it also seems reasonable to conclude that the UK probably has an autism epidemic, too, but just hasn't yet realised it.
Ÿ Dr Bernard Rimland of the US Autism Research Institute, San Diego: "Some supposed experts will tell you that the (US) increase reflects only greater awareness. That is nonsense. Any paediatrician, teacher or school official with 20 years experience will confirm there is a real increase, and the numbers are huge and growing".
As in the UK, health officials in the US have tried to explain away these increases as being the result of greater awareness, better recognition and broader diagnostic definition. Doubtless these play a part, but the authorities seem to want to use these factors to explain all the increase, without having any hard evidence to support their stance.
In April 2000, giving evidence to the Government Reform Committee hearings into autism's increase, Dr. Coleen Boyle, Associate Director for Science and Public Health at the Center for Disease Control, stated that UK rates in 1966 had been 4 to 5 per 10,000 (1 in 2,500-2,000). Studies from outside the US since 1985 had indicated 12 per 10,000 (1 in 833). Recent studies had been higher still. There had been only two population-based studies in the US, both in the 1980s, indicating prevalence of 1.2 to 3.3 per 10,000 (1 in 8333 to 1 in 3030).
Two years on, giving evidence to the same Congressional committee, Dr. Coleen Boyle acknowledged the case of Brick Township New Jersey, where the CDC had found a rate of ASD of 6.7 per 1,000 (note: per ONE thousand), or 1 in 149. She stated that the previously-accepted background rate was 1-2 per 1,000 (comment - but this does not square with her evidence in the year-2000 Washington hearings?). She stated "We cannot determine whether rates are increasing or not, because we do not have comparable data from earlier years".
But the thrust of her earlier comments implied that, even if increases were demonstrated, this was down to better awareness etc., and at no point did she appear to confront the possibility that increases were real, and then confront the (very difficult) question, "What was causing the increase?". The CDC strategy seems to be to cast doubt upon the increase, rather than seek the cause. The CDC strategy might be summed up as follows:
Ÿ Cast doubt upon the accuracy of the data, and thus draw attention, and the focus of debate, away from the cause of the increase and towards the data issue
Ÿ Stress the need for better data (which no one would argue against)
Ÿ Announce new comprehensive data-gathering exercises, which will take time.
By early 2003, evidence that increases were real was beginning to accumulate - see next main section.
23. Autism Elsewhere
Information on autism in Canada does not appear to be anything like as comprehensive as that in the US, but press reports are indicating a recent increase. In May 2002, a study by the Ontario government health ministry indicated that numbers were increasing sharply, with 800 children younger than six years of age being newly diagnosed during 1998. This represented a 53% increase over numbers diagnosed two years earlier.
The Ontario government study also found that 2,863 children younger than seven were diagnosed with autism between 1991 and 1998. The study was not released until the efforts of a parent, Professor Marianna Ofner-Agostini of the University of Toronto, forced the issue.
The issue is now being debated in other developed countries elsewhere in the world. A New Zealand doctor, Dr. Mike Godfrey, wrote to the UK Scotsman newspaper in early 2002 as follows: "I have so far analysed 866 children's histories, with 260 being unvaccinated. There are no cases of autism, epilepsy or Crohns Disease and only a handful of other diseases in this latter (unvaccinated) group. There are 16 autistics, 12 epileptics, 8 cases of Crohns, plus cases of other illnesses, in the vaccinated 606 children."
PART C:
EVIDENCE THAT INCREASES ARE REAL
24: California
California has probably the most useful and detailed autism data in the world, going back to 1970. Trends monitored there have a potential worldwide significance.
Ÿ The rise in autism was first highlighted by a report Changes in the Population of Persons With Autism and Pervasive Developmental Disorders in California's Developmental Services System, 1987 through 1998 - A Report to the Legislature, tabled on March 1st 1998 by the Department of Developmental Services, Sacramento, California Health and Human Services Agency.
Ÿ Department of Developmental Services data, released at the start of 2002, shows that a record number of professionally-diagnosed DSM-IV criteria autism cases, 2,725 cases, entered the State system during 2001.
Ÿ This year-2001 number represents a 20% increase over the year 2000, itself a record.
Ÿ In 2001, there were more cases of level-one autism in California than in 1994, 1995 and 1996 combined.
Ÿ Historically, autism made up 3% of childhood disability in the State Developmental Services system. It now comprises 35% of the total.
Ÿ Two out of three persons with autism in California's child-developmental system are now young children between the ages of 3 and 13. Eight out of ten persons with autism have been born since 1980 (1980 was the year that California mandated the full complement of childhood vaccines as a condition of school entry. MMR was also introduced in California 1979-80).
Ÿ California now has 16,802 persons with level-one autism in its Developmental Services system.
Ÿ The total intake for the three years 1999-2001 was 6,596. This compares with a total intake for the twenty-five years 1970-1995 of 6,527 cases.
This does not include children with persistent developmental disorder, non-specific (NOS) developmental delays, Asperger's or and other autistic spectrum disorder - it is therefore the tightest definition of the severe-case numbers.
Statistics on autism in the individual regional centres in California, run by the state Department of Developmental Services, also show a sharp rise in the period 1998-2002:
| (regional centre) | At 7th Jan 1998 | At 3rd Jan 2002 | Increase % |
| Alta | 400 | 683 | 71% |
| Central Valley | 150 | 361 | 141% |
| East Bay | 606 | 1,087 | 79% |
| E. Los Angeles | 443 | 976 | 120% |
| Far Northern | 125 | 217 | 74% |
| Golden Gate | 371 | 499 | 35% |
| Harbor | 639 | 1,113 | 74% |
| Inland | 568 | 1,195 | 110% |
| Kern | 141 | 262 | 86% |
| Lanterman | 418 | 842 | 101% |
| North Bay | 215 | 350 | 63% |
| N. Los Angeles | 742 | 1,746 | 135% |
| Orange | 670 | 1,621 | 142% |
| Redwood Coast | 76 | 103 | 36% |
| San Andreas | 360 | 666 | 85% |
| San Diego | 609 | 1,186 | 95% |
| San Gab/Pomona | 581 | 937 | 61% |
| S. Central LA | 549 | 874 | 59% |
| Tri-Counties | 352 | 725 | 106% |
| Valley Mountain | 153 | 373 | 144% |
| Westside | 613 | 986 | 61% |
| (Statewide Total) | 8,781 | 16,802 | 91% |
Since this data was released, further rises have been recorded:
Ÿ The quarterly data reported from the California Department of Developmental Services for the period April-July 2002 added a further 846 new children with level-one DSM-IV autism to the Department's caseload. These 846 cases represented a new record high in the 31-year history of the system.
Ÿ The increase was 18% higher than for the comparable quarter in the previous year. Autism now constitutes 40% of all developmental-disability intake for the California system.
Ÿ On average, nine new cases are being added to the system, every day, seven days a week.
Ÿ To set the 2002 figures in context, over the 28 years from 1971, to 1999, California produced 10,206 cases. In the subsequent three and a half years, 1999 to mid-2002, a further 8,554 new cases were added.
Comment: the above suggests a major rise in autism incidence in California.
25. The MIND Study, California
Following mounting concern at the apparent steep increase in autism in California, an urgent study was launched by the MIND Institute. Its findings were released on 17th October 2002, and appear to finally confirm (but see other contradicting studies in the following section) that autism has risen steeply.
The study was led by Dr. Robert Byrd, whose team had previously enrolled 684 Californian children who were receiving services from one of the Department of Developmental Services regional centers. Byrd's team systematically gathered information for children in two age groups, 7-9 year olds, and 17-19 year olds. These were drawn from families of 375 children with a diagnosis of full-syndrome autism, and families of 309 children with a diagnosis of mental retardation without full- syndrome autism.
The study findings were that:
Ÿ The unprecedented increase in autism in California is real and cannot be explained away by artificial factors such as misclassification and criteria changes. Autism is on the rise in California and the study team does not know why
Ÿ The observed increase cannot be explained by a loosening in the criteria
Ÿ Some children reported with mental retardation and not autism did meet criteria for autism, but this misclassification does not appear to have changed over time
Ÿ Because more than 90% of the children in the survey are native to California, major migration of children into California does not contribute significantly to the increase in autism
Ÿ A diagnosis of mental retardation associated with autism had declined significantly between the two age groups studied.
Ÿ The percentage of parent-reported regression (loss of milestones) does not differ between the two age groups studied
Ÿ Gastrointestinal symptoms, including constipation and vomiting, in the first fifteen months are more commonly reported by parents in the younger group
Comment: the above study appears to offer firm evidence of a major rise in prevalence.
26. New Jersey
Data on autism in New Jersey, recorded by the IDEA system for individuals with disabilities who require special education, suggest that there is a vast preponderance of cases amongst children/young people ages 6-21 amongst the youngest ages. The following figures relate to the position as at 1st January 2002:
| age | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 |
| nos | 514 | 505 | 465 | 439 | 360 | 257 | 208 | 165 |
| age | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 |
| nos | 145 | 124 | 81 | 73 | 58 | 63 | 30 | 14 |
The total number of cases is 3,501. This equates to an average of 219 for each age-year. For ages 11 and under, the number exceeds this average, and for ages 12 and over, it is less than this average.
The youngest three years average out at 495 cases. The oldest three years average out at 36 cases. The average of the youngest years is about 14 times that of the oldest three years.
In an article published by the US Autism Autoimmunity Project at the end of December 2002, Dr. Ed Yazbak set out the evidence for there having been a huge rise in autism in Rhode Island, New Jersey:
Ÿ The Special Education Census published yearly by the Rhode Island Department of Education listed 14 categories of primary disability, by school district. Two categories, autism, and behavioural disorders, had risen sharply.
Ÿ Autism had increased by 1,115% (one thousand, one hundred and fifteen per cent) between 1994 and 2002 in Rhode Island schools. On 30th June 1994, there had been 41 students at the schools with a diagnosis of autism. By June 30th 2002, that number stood at 498.
Ÿ The more restrictive diagnostic criteria of DSM IV had been used, exclusively, since 1994, and had remained unchanged. Rhode Island has one main diagnostic center, one paediatric psychiatric hospital and very few paediatric neurologists, so consistency in application of diagnostic criteria would be high.
Comment: the above seems to confirm that the recent very steep rises in California are also being witnessed elsewhere in the US.
27. Atlanta Study, Prevalence of Autism in a US Metropolitan Area, by Yeargin- Allsopp, Rice et al, published in the Journal of the American Medical Association, 2003, Jan 1st, 289: (1): 49-55
This study was at last an acknowledgment at the US Center for Disease Control & Prevention that autism was at a higher real level than two decades ago. Its conclusions directly undermined the evidence of one of its participants, Dr. Coleen Boyle, to the US House of Representatives Government Reform Committee, only a short time earlier, that autism was a very rare condition.
Ÿ The objective of the study was to determine the prevalence of autism among children in a major US metropolitan area, and to describe the characteristics of the study population.
Ÿ The study looked at children aged 3 to 10 years in the counties of metropolitan Atlanta, in 1996. Cases were identified through screening and abstracting records at multiple medical and educational sources, with case status determined by expert review.
Ÿ The results were that 987 children were identified, displaying behaviour consistent with DSM-IV criteria for autistic disorder, PDD-NOS or Asperger disorder.
Ÿ The prevalence for autism was found to be 34 cases per 10,000
Ÿ The conclusion was that the rate of autism found was higher than the rates from studies conducted in the US during the 1980s and 1990s, but was consistent with those of more recent studies.
Comment: this study, too, supports the view that autism has greatly increased. The study is notable for being a CDC-sponsored study, using CDC personnel.
PART D:
REVIEWS QUESTIONING THE AUTISM EPIDEMIC
Despite the evidence that autism has increased very greatly since the 1970s and early 1980s, several researchers maintain that this is not the case.
28. Paper by Fombonne, Medical Research Council Child Psychiatry Unit and Institute of Psychiatry, Is There An Epidemic of Autism?, Pediatrics, January 2001
At the end of January 2001, a paper, "Is There An Epidemic of Autism?" was published by Dr. Eric Fombonne. The paper sought to deny that autism had really increased, and criticised the "poor research methodology" of Dr. Andrew Wakefield, and said "There is no need to raise false alarms on putative epidemics nor to practice poor science....."
Ÿ Fombonne criticises the California increase on the basis of in-migration, possible changes within the population make-up, the change from DSM-III to DSM-IIIR in 1987, the introduction of diagnostic categories for Asperger, Rett and childhood disintegrative disorder in DSM-IV in 1994, the effects of earlier diagnosis adding to the totals, and other factors.
Ÿ His most useful conclusion is that "we simply lack good data". He raises doubts about the apparent epidemic, but is then unable to refute it either.
In an excellent FEAT (parents' group) critique (8th Feb 2001), Mark Blaxill goes carefully through Fombonne's previous work and argues that Fombonne has become inconsistent. He points out key flaws in Fombonne's previous work, and criticises his criticisms of the California data and his scientifically-unsupported assertions
29. Paper by Lorna Wing, Centre for Social & Communication Disorders, Elliot House, Bromley, Kent, UK and David Potter, UK National Autistic Society, The Epidemiology of Autistic Spectrum Disorders: Is The Prevalence Rising?, 2002
This paper noted that:
Ÿ For decades after Kanner's original paper in 1943, autism was generally considered to be a rare condition with a prevalence of around 2 to 4 cases per 10,000 children. Then in the late 1990s, prevalence rates of up to 60 cases per 10,000 for autism, and even more for the whole ASD spectrum, were reported.
Ÿ Reasons for this included changes in diagnostic criteria, development of the concept of the wide autistic spectrum, different methods used in studies, growing awareness and knowledge amongst parents and professionals, the development of specialist services, and the possibility of a true increase in numbers.
The paper argued that not one of the possible environmental causes, including MMR, had been confirmed by independent scientific investigation
The paper maintained that there was "strong" evidence that complex genetic factors played a major role in aetiology (Comment: this point and the one above seemed to be treated as "either/or" explanations rather than in combination)
In direct contrast with the 2002 California paper, this paper concluded that "the evidence suggests that the majority, if not all, of the reported rise in incidence and prevalence is due to changes in diagnostic criteria and increased awareness and recognition of autistic spectrum disorders. Whether there is also a genuine rise in incidence remains an open question".
30. Position of Dr. Bryna S. Siegal, Director, Pervasive Developmental Disorders Center, University of California at San Francisco, 2002
The August 2002 issue of Paediatric News carried a report by Sherry Boschert, about the position of Dr. Bryna S. Siegal of California, expressed at a meeting on developmental disabilities sponsored by the University of California at San Francisco. Dr. Siegal's view is that:
Ÿ Prevalence in autism in California increased from 5 per 10,000 in 1987 to 15 per 10,000 in 1994, yet during the same time, diagnosis of mental retardation declined by a similar amount, dropping the State prevalence of mental retardation from about 27 per 10,000 to around 18 per 10,000.
Ÿ Changing social attitudes have shifted stigma away from autism and onto mental retardation
Ÿ Autism is partly now preferred because it is associated with a higher level of State services. Dr. Siegal claims that many letters from parents actively seek a diagnosis of autism
Ÿ These are not the only factors fuelling what she describes as an "illusory" epidemic of autism. The inclusion of the diagnosis of pervasive developmental disability into the former DSM-III classification in 1980, creating DSM-IIIR (or III-revised) resulted in autism rising by one-third. In 1994, the creation of DSM-IV, which included Aspergers cases, further increased the numbers.
Comment: these views have been strongly contradicted by:
Ÿ The views of parents, professionals and others, who testify that autism is now being seen in unprecedented numbers
Ÿ The point that the autism of the past largely comprised children who were autistic from birth or from a very young age, and that the "new variant" regressive autism was apparently largely unseen and unreported until the late 1980s, and that it is extremely unlikely that dramatic regression and loss of milestones would have been missed in the past
Ÿ Detailed research carried out by Dr. Robert S. Byrd in late 2002 (reported elsewhere in this note), in California, finds that the apparent increase in autism is real, and not ascribable to reassignment from other categories
31. Study by Croen, Grether, Hoogstrate and Selvin for the California Department of Health Services, July 2002
The authors conducted a population- based study of eight successive birth cohorts to examine the degree to which improvements in detection and changes in diagnosis have contributed to the observed increase in autism prevalence. Children born in 1987-1994 who had autism were identified from State registries. To evaluate the role of diagnostic substitution (re- assignment from other categories), trends in prevalence of mental retardation without autism were also investigated.
Ÿ A total of 5,038 children with full-syndrome autism were identified from 4,590,333 births, giving a prevalence of 11 per 10,000
Ÿ During the study period, prevalence of autism increased from 5.8 per 10,000 to 14.9 per 10,000
Ÿ During the same period, the prevalence of mental retardation without autism decreased from 28.8 per 10,000 to 19.5 per 10,000.
Ÿ The data, in the view of the researchers, suggests that improvements in detection and changes in diagnosis accounts for the observed increase in autism. However, they also conclude: "Whether there has also been a true increase in incidence is not known".
Comment: this report backs the views of Dr. Siegal (see above) and Dr. Fombonne (see below), but contradicts the study by Dr. Byrd (see elsewhere). The authors also acknowledge that there study cannot rule out that there has been a real increase. The criticisms applied to Dr. Siegal's work also apply here.
32. Fombonne, editorial, Journal of the American Medical Association, January 1st 2003 Vol 289, No.1 49
At the start of 2003, Dr. Eric Fombonne wrote an editorial in the Journal of the American Medical association that appeared to acknowledge that there had been some real increase in autism, but which also attempted to explain this away to as great a degree as possible through the usual recourse to references to better awareness, less restrictive criteria and a greater willingness to diagnose.
Fombonne's key points were that:
Ÿ That the prevalence rate of 34 per 10,000 (1 in 294) was likely to actually be an underestimate, because high-functioning autism cases were likely to have been missed.
Ÿ The lower reported prevalence in 3- and 4-year olds might reflect lower sensitivity of case identification for disorders, which were often diagnosed later
Ÿ There was an unexpected decrease in prevalence amongst 9- and 10-year olds. Fombonne dismisses the idea that this might imply that the younger the birth cohort, the greater the level of autism as being "biologically implausible". Yet this is open to obvious question - what if an external factor had altered during this time? Fombonne does not address this possibility.
Fombonne concluded that a rate of 41-45 per 10,000 (1 in 222) might be a more accurate rate of prevalence. He noted in his editorial that other studies suggested rates of 60 per 10,000 when pervasive developmental disorder-not otherwise specified (PDD-NOS) and Aspergers syndrome were taken account of.
He then addressed the issue as to whether the prevalence of autistic spectrum disorder (ASD) had increased over time. His benchmark was the 1970s Wing and Gould study in Camberwell, London, which pointed to a rate of 20 per 10,000 for severe-impairment cases. Other earlier studies had point to rates of 4 or 5 per 10,000, and more recent studies cited by Fombonne pointed to rates of more than 10 per 10,000. Fombonne's conclusion was that the most recent rates of prevalence were three or four times higher than 30 years ago.
Fombonne, seemingly searching for an uncontroversial explanation for any increase, then examined whether this increase implied a broadening of criteria and improved methods of case-finding during studies. He pointed to what he described as the "major" changes in criteria:
Ÿ Diagnostic and Statistical Manual of Mental Disorders, Third Edition (DSM-III), 1980
Ÿ DSM Revised Third Edition (DSM IIIR), 1987
Ÿ DSM Fourth Edition (DSM IV), 1994.
He argued that there was strong evidence that differences in methods for case finding could account for a "huge" proportion of the variability of prevalence estimates between surveys. Referral rates were also unreliable, due to confounding factors. This, and other factors, he concluded, combined to offer "good" evidence to support the contention that higher rates of prevalence reflected changes in diagnostic practice, improved identification and availability of services. The hypothesis of an increasing trend in the incidence of autism could not, in his view, be fully tested because of the inadequacy of studies to date. Fombonne dismissed any association with MMR (citing his own study work and studies by Madsen and by Taylor and Miller as proof), and dismissed evidence of any connection with thiomersal as being "weak".
Fombonne was also quoted in the New York Times of 31st December 2002 as stating: "No strong candidate environmental exposures have been identified.....Claims of an association with MMR have not been borne out by recent studies, and evidence for causal association with other exposures such as mercury-containing vaccines is weak".
The study being commented on by Fombonne was that by Dr. Marshalyn Yeargin-Allsop et al, detailed earlier.
Comment: the editorial by Fombonne offers no hard evidence against a vaccine/autism link, and, whilst offering some arguments in favour of questioning the precise scale of the apparent major rise in autism prevalence, fails to demolish the central assertion of many parents, that autism has grown immensely in a couple of decades. No alternative explanations for the rise are offered by the Fombonne editorial.
PART E
STUDIES THAT HAVE BEEN USED TO DENY AN MMR/AUTISM LINK
This section deals with the numerous recent official studies and reviews, many in the UK but some in the US or elsewhere, that "prove" there is no connection between autism and vaccination.
As will be seen, all when scrutinised critically are actually either irrelevant, inconclusive, or are seriously methodologically questionable.
Ÿ The UK Government's advice on MMR and autism comes from the DoH, the Medicines Control Agency (MCA), the Committee on Safety of Medicines (CSM) and the Joint Committee on Vaccination and Immunisation (JCVI). These bodies are closely intertwined, and have based their position on a barely more than a handful of studies. Further advice has come from the Medical Research Council.
Ÿ Much of the focus has been upon the need maintain public confidence in MMR to prevent communicable diseases, rather than the need to investigate specific cases of alleged damage.
Ÿ The studies are also of random groups of children, but not of the actual children reported by parents as damaged by MMR.
Ÿ Finally, the Department of Health has implied in the past that the evidence for a link between MMR and regressive autism has come from only one team of researchers, which is not factually correct. However, the very same criticism can be levelled at the "anti-link" camp. A significant proportion of the studies below only comes from a very small number of sources, some very close to the Department of Health itself.
33. Stokes et al Paper, Trivalent Combined Measles Mumps Rubella Vaccine, Journal of the American Medical Association, 4th October 1971
This paper, by Stokes, Weibel, Villarejos, Jorge, Arguedas, Buynak and Hilleman, has assumed more importance recently (see later Wakefield/Watson/Shattock debate section).
Ÿ The paper stated that triple vaccines were desirable to simplify administration, reduce costs and minimise visits (my emphasis). There was no mention of greater effectiveness, or inherent drawbacks with single vaccines.
Ÿ There were three trials, firstly of 30 children in Philadelphia, then of 214 children in Philadelphia, then of 440 children in rural Costa Rica and San Salvador, total 684 but (note) of very different economic and geographical backgrounds.
Ÿ The mean ages of children in the three trials was 1.1, 1.5 and 3.0 years. Note that the present age of receiving MMR is about 14 months, and therefore the vast majority of the trial children were significantly older than today's UK MMR recipients. Some 64% were also not from Western social/health backgrounds.
Ÿ The 30 children's parents were given report cards for recording temperatures for 28 days, and queried at six to nine weeks. For the 214 child-cohort and the 440 child-cohort trials, follow-up was 28 days. The parents were instructed to notify any significant illnesses during the 28-day period, and were queried at the second bleeding, six to nine weeks after vaccination - but the implication is that this query may have covered the 28-day interval, not longer.
Ÿ The study noted that "the fifth to twelfth day after vaccination is the critical time period for occurrence of the expected low incidence of febrile reaction", also that the significance of the difference between vaccinees and controls in terms of miscellaneous subsequent complaints (gastroenteritis included) was "doubtful" (though it was actually very marked in the study tables, up to 18/228 of vaccinees with gastroenteritis, compared with at most 3/106 of controls)
Ÿ At no point in the study was autism mentioned as a risk-factor or an actual outcome. Clearly, the possibility was not even considered. The study noted the lack of arthritis and arthralgia.
Overall verdict: this study is not relevant to disproving an MMR/autism link
34: Study of Twins By Peltola and Heinonen, Frequency of True Adverse Reactions to MMR Vaccine; A Double-Blind Placebo-Controlled Trial in Twins, National Public Health Institute and Children's Hospital, University of Helsinki, Finland, published Lancet, April 26th 1986
This study sought to check levels of adverse reactions following MMR. MMR was introduced into Finland in 1982, being administered at 14-18 months and at 6 years, using Merck Sharp Dohme Viravac.
Ÿ The study was a double-blind crossover study involving 581 twins. The vaccines were administered blind, but one twin of each pair first received active MMR, then three weeks later, received a placebo. The other twin was given the placebo first, then three weeks later received MMR.
Ÿ Each twin was given a colour coded questionnaire to be completed daily by parents, for 21 days after the injections.
In theory, this should have provided a foolproof test of how reactive MMR was. However, the study completely founders on:
Ÿ the issue of the potential time-delay between receipt of MMR and any possible gradual degeneration into autism. If such a delay could exceed 21 days, then the study would have missed it as an adverse reaction
Ÿ Secondly, the linking of autism/developmental delays with MMR, or indeed any other vaccine. Parents in 1982, or indeed until about mid-1997, were not linking MMR with autism. It is extremely unlikely that regressive autism would have been connected, in the minds of either parents or the study authors, with MMR back in 1982. Virtually no literature or press reports had appeared on the issue.
Ÿ As with the original safety trials of MMR (see later papers), this study was not designed to verify whether rare and complex adverse events might follow months or years after MMR.
Ÿ The study only looked at one brand of MMR. As subsequently transpired, some brands of MMR used in the UK and elsewhere had a less satisfactory safety record than others, and (in the UK) were withdrawn at very short notice in 1992. A study with Viravac cannot be used to give safety clearance to other brands if the brands are found to have been variable.
Ÿ A further criticism is that the study is still quite small in relation to rare events. It involved 581 twins. All other things being equal, if a rare adverse outcome occurred at a rate of 1 in 2 x 582, or less frequently, this study would not have found it.
The authors did actually acknowledge this, stating:
Ÿ "The study was designed to explore relatively common symptoms and signs occurring after the vaccination" (they mean, "within 21 days of"), and
Ÿ "Rare reactions due to the MMR vaccine cannot be studied with this small sample".
It is therefore suggested that this study, regarded as the "gold standard" by the exponents of MMR, offers no evidence for or against an MMR/autism link; it is clearly irrelevant. Overall verdict: this study is not relevant to disproving an MMR/autism link
35: Study by Miller, Miller, Rowe et al, Surveillance of Symptoms Following MMR Vaccine in Children, The Practitioner, Vol 233, 8th January 1989
This paper was to report the incidence and severity of clinical reactions before the start of the UK national MMR programme. MMR was offered to 10,000 children in three districts in the UK, with a post-vaccination follow-up of every child.
Two types of MMR were introduced, Immravax in Somerset, England, and Pluserix in Fife, Scotland, and North Hertfordshire, near London. Both vaccines contained Schwarz measles and Urabe 9 mumps vaccine, and both later had to be withdrawn in 1992 for safety reasons, in connection with risks of aseptic meningitis. These risks were not detected by this study.
The study found that:
Ÿ Of the 7,247 children aged 1-2 years, 38% had either no symptoms or symptoms for only one day
Ÿ 18 had convulsions. Fifteen were admitted to hospital.
Ÿ Of the children aged 4-5 years, 61% had either no symptoms or symptoms for one day. There were no convulsions and no hospital admissions.
Ÿ Follow up was for 21 days. However, 114 children were followed up through diary records for a further 21 days, total 42 days.
Ÿ Comparison of symptoms of children after MMR was made against symptoms of children after measles vaccination - not unvaccinated children.
Ÿ The study concluded that symptoms reported after MMR appeared to be similar in nature, frequency, time of onset, and duration, to those recorded in earlier studies after monovalent measles vaccine
Comment: as with the original safety trials of MMR, follow-up was extremely short and only immediate/near-immediate reactions noted. The study did not look at autism, but effectively cleared the way for MMR's general introduction into the UK. It is noteworthy that the study was co-authored by Dr. Elizabeth Miller, who subsequently authored or co-authored several of the studies that have been used as "proof" that there is no MMR/autism link. It is also noteworthy that, as noted, this study missed the aseptic meningitis problem of MMR, and that the brands of MMR with Urabe strain mumps virus subsequently had to be withdrawn, in 1992, at extremely short notice.
Overall verdict: this study fails to disprove an MMR/autism link
36. Gillberg Study, Sweden, Is Autism More Common Than Ten Years Ago?, British Journal of Psychiatry, 1991, 158; 403-409
The paper reported a study in Sweden by Gillberg et al, 1991. It has been partially updated since (see below).
Ÿ Gillberg looked at tiny sample of autistic children (55 of typical autism, just 19 of atypical autism), in Goteburg and Bohuslan. The study, actually a mish-mash of three studies with differing criteria, does not mention vaccination, does not state the coverage of MMR, does not include data on uptake or demographic factors, and is therefore irrelevant to the MMR/autism debate.
Ÿ It had tracked down cases of autism unscientifically, by word of mouth, doctors etc., then allocated them by d.o.b. to "pre-MMR" and "post-MMR" eras
Ÿ The study's case-selection being a few cases out either way would neutralise or completely reverse the findings of the study.
Ÿ The paper does acknowledge that the rate of autism has increased but "explains" this through changes in population structure and "better diagnosis".
Overall verdict: this study offers little evidence that MMR does not cause autism, particularly as it is so small.
37. Paper by Gillberg and Heijbel, Commentaries, Autism, Vol 2 (4) 423-430, 1998
This further paper by Gillberg was published following the appearance of the Wakefield et al "Early Report" paper in The Lancet in early 1998.
Gillberg and Heijbel stated that they had re-analysed the data from their population study of autism performed in the late 1980s and published in 1991 (as above). The children in that study (n = 55) had been born in the ten-year period 1975-84. The authors claimed that as MMR was introduced in Sweden for 18-month-old children in 1982, with coverage increasing rapidly to 90%. The authors then argued that if there was an MMR/autism link, then children born from July 1980 onwards (i.e. The post-MMR generation) would be expected to be at increased risk. The 55 children were therefore divided into 34 (62%) pre-MMR and 21 (38%) post-MMR.
The authors then argued that had there been a strong effect of MMR, they could have expected more than 45% of the 55 cases of autistic children to have fallen into the post-MMR group. As this was not the case, then their study did not support the hypothesis of an association between MMR and autism
The authors also again claimed that in their parallel study of 19 atypical autism cases, there would have been a similar effect, and therefore that again there was no support for an association.
Overall verdict: as with the original study, these numbers were so small as to render this study, and its conclusions, as virtually without value in the context of proving/disproving an MMR/autism link. Statistical/epidemiological studies based upon cohorts numbering 55 and 19 cases are far too small. It is extraordinary that the UK Department of Health was using this study in the late 1990s to "disprove" the suggested association.
38. Letter by Dr. Eric Fombonne, Inflammatory Bowel Disease and Autism, Pediatrics, March 28th 1998
This letter set out two studies that attempted to prove that there was no connection between inflammatory bowel disease/Crohn's disease and autism. The first study looked at UK clinical data collected by the Child & Adolescent Psychiatric Services of the Maudslay Hospital, London.
Ÿ For ASD, three diagnostic groups were examined, autism, atypical autism including disintegrative disorder, and pervasive developmental disorder
Ÿ Medical disorders were coded for a 25-year period, including Crohn's and ulcerative colitis, for 8889 patients.
Ÿ Of the 8889 patients, 987 were born in 1987 or later, and were therefore most likely to have been exposed to MMR. Of these, 201 had ASD.
Ÿ Of the 8889 children, only two had Crohn's, and both were non-autistic. None had ulcerative colitis.
For the second study, a similar approach was undertaken. Fombonne surveyed medical, behavioural and intellectual disabilities amongst 6100 French children.
Ÿ He found 174 cases with autism.
Ÿ One child of the 6100 had Crohn's, and one had ulcerative colitis. Neither were autistic
Ÿ The conclusion that Fombonne drew was that these data provide no support for the hypothesis of an association between IBD and autism.
Overall verdict: neither of these studies offer any evidence to disprove an MMR/autism link.
39. UK Committee On Safety of Medicines Study, Report of the Working Party on MMR Vaccine, Committee on Safety of Medicines, June 1999
This study looked at the medical records of some of the children who are now taking High Court action. Their details were provided by their lawyers.
The study admitted:
Ÿ Information on the children was extremely variable in quality and completeness
Ÿ It was "difficult" to draw conclusions about any causal association (verbatim quote: "the information evaluated has important intrinsic limitations as regards assessing whether the vaccines are or are not causally associated with the adverse effects")
Ÿ It was not feasible to review the less common adverse side effects
The study was effectively run as knockout competition. Each case had to pass four hurdles (all four) to be counted as being caused by MMR. The four hurdles were: (1) have either the diagnosis or clinically relevant signs/symptoms been confirmed medically? (2) was the onset of the possible adverse effect within six weeks of immunisation with MMR? (3) was there history prior to immunisation relevant to the possible adverse effect? (4) was there evidence of other causes for the possible adverse effect?
Ÿ Six weeks after immunisation was chosen as a cut-off point for a close temporal association because (quote) "this is the maximum period in which viral replication can be detected after immunisation". But this probably missed many cases, and is arbitrary. The Spitzer, Aitken et al study (see later) renders this six-week limit as irrelevant.
Ÿ At every stage, the study looked for other "causes" to explain-away the cases, and took every opportunity to ascribe cases to these "causes". In most cases, it was assumed at every stage without scientific justification that autism was "caused" by other factor rather than MMR. But it is not known what causes autism. Therefore there is a gross study bias, and the study rests upon unscientific assumptions.
Ÿ The other assumed "causes" were the child's previous medical history, comprising having a parent/sibling with speech or behavioural problems, an obstetric history of pregnancy complications (these, alone, were not considered as "causes"), signs/symptoms of encephalopathy, a head circumferences larger than the 97th percentile, or history of unspecified viral illnesses, bronchiolitis, rubella, measles, or a minor head injury.
Ÿ The study eventually only looked at 92 cases of autism in detail (plus 15 Crohn's), and was left with a residue of 8 autism cases and four of the Crohn's it could not "explain" away. These were then just set aside, without explanation.
Ÿ What the study did was to introduce so many extraneous considerations, and accord these such an importance, that hardly any case with sufficiently-clear documentation remained to survive the appraisal process. This eliminated almost all cases. The study then appears to have then simply set aside the residue.
Ÿ The study text commented that (quote) "it was impossible to prove or refute the suggested associations between MMR vaccine and autism or inflammatory bowel disease because of the nature of the information.". This would seem to inevitably render the study as inconclusive. But the study's conclusions did not reflect this sentence.
Ÿ The wording of the final conclusion left a small exit-route for any possible future U-turn: ""On the basis of all the available evidence, the demonstrated benefits of MMR or MR vaccines far outweigh any possible risks" (my emphasis).
Ÿ The DoH's press release 0342 of 1999 spun the study's conclusions further - "Two New Independent Studies Have Not Found A Link Between MMR Vaccination And Autism"
Note: this is the only study to date to have both looked at the actual children reported to have been damaged and to have "cleared" MMR. But as the above criticisms show, the study was actually self-admittedly inconclusive. It also failed to medically examine the actual children.
Overall verdict: this study does not disprove an MMR/autism link.
40. Paper by Taylor, Miller, Farrington et al, Autism and Measles Mumps Rubella Vaccine: No Evidence for a Causal Association, Lancet 1999, 353, 2026- 9
The study, designed by Dr Elizabeth Miller of the Public Health Laboratory Service, was wholly inconclusive, but has been widely presented as conclusive proof of the absence of any link between MMR and autism.
Ÿ It only looked at 498 cases, far too small a sample for a robust statistical (case-series analysis) test. The study attempted to track-down children through special schools and local authority special needs registers - a method that is open to question, as it probably misses many cases. The study describes itself as "a large regional sample", but it was actually very small.
Ÿ Taylor, Miller found a steep increase in autism, ("There was a steady increase in cases by year of birth"), but did not explain it.
Ÿ Also, the study looked for a time-clustering of parental concern six months after MMR, found it, but then dismissed it unconvincingly by saying it was "related to the difficulty of defining precisely the onset of symptoms". But this method, of precisely identifying a date, was meant to be the very basis of their study.
Ÿ Also, the study did not include in its post-MMR numbers those children born 1986-87 who later received it, nor those 2/3/4 year olds who had MMR at this older age.
Ÿ It also missed children who had single vaccines, then MMR later. It not only misses these from "post-MMR" numbers, but added them to its pre-MMR numbers. The whole study is thereby compromised. The authors have since sought to clarify this in correspondence in The Lancet, but unconvincingly.
Ÿ Autism is sometimes not diagnosed for years after. It is very difficult to pin down an actual "date" of diagnosis, and many children don't receive any formal diagnosis anyway (contact National Autistic Society, which did a study on this, tel 0207 833 2299). The Taylor Miller study doesn't recognise this.
Ÿ The study seems to have been designed to clear MMR, not to test whether there is a link with autism. The study struggles, and fails, to disprove a link.
Ÿ Also, the study is described by the UK DoH as "independent". But Taylor was co-author of a 1988 paper clearing the safety of triple vaccines, Miller was described in Daily Express press reports of 1/01 as "a colleague of Dr David Salisbury" (head of the DoH Immunisation & Communicable Diseases Branch, which runs the MMR programme), and the study was funded by the UK Medicines Control Agency, a satellite of the DoH.
Ÿ The authors have been repeatedly challenged by other researchers to release their raw data but have refused. Yvette Cooper, the UK Minister for Public Health, has backed up their refusal.
Overall verdict: despite its claims, this study cannot be taken as proof of there being no MMR/autism link, due to its apparent serious methodological flaws.
(Note: this study has been claimed by the UK Medical Research Council to represent "strong positive evidence" of there being no MMR/autism link)
41. Paper by Miller and Farrington to US Government Reform Committee Hearings, Written Testimony to the Congress of the United States Committee on Government Reform Hearing On The Challenges of Autism - Why The Increased Rates, April 2001
In their submission to the US House of Representatives Committee on Government Reform Hearing, which was investigating increases in autism and possible links with vaccination, Miller and Taylor re-stated:
Ÿ "Our conclusion, based on the findings of our study, is that there is no evidence of a causal association between MMR and autism".
Ÿ "The case series method has a proven track record with respect to identifying and measuring a risk of adverse events after various vaccines".
Ÿ "In our study, we showed that the increase in the prevalence of diagnosed autism in recent birth cohorts occurred during a time when the coverage of MMR vaccine in the same cohorts has been constant. The rise cannot therefore be related to the use of MMR vaccine."
Ÿ "There is no credible epidemiological evidence to support the view that measles vaccination is a risk factor for Crohn's disease or any other inflammatory bowel disorder".
However, as explained in the section covering the original paper by Miller, Taylor and Farrington, there are major questions over the methodology of this paper; these, of course, can also be applied to Miller and Farrington's paper to the Government Reform Committee.
42. Patja, Peltola et al Study, Serious Events Rarely Related to MMR Vaccine: Natural Diseases Outweigh Risks, Pediatric Infectious Disease Journal, 2000;19; 1127-1134 (December)
This Finnish study, usually referred to as the Peltola study, concluded that serious events rarely were related to MMR. The study was initiated in 1982, when MMR was introduced. A nationwide surveillance system was set up to detect serious adverse events, reviewing patients' clinical records and where taken, serum samples. However, the study relied on passive surveillance - a fatal flaw - and only followed up acute adverse events - a further fatal flaw.
According to the report,
Ÿ 173 potentially serious adverse reactions were claimed to have been caused by MMR, out of almost 3 million doses.
Ÿ There were 77 neurologic reactions, 77 allergic reactions, 22 miscellaneous reactions and one death.
Ÿ Some 45% of these reactions were dismissed by the study as probably caused or contributed by other factors.
Ÿ Peltola admitted on BBC Radio 4 on 13/1/01 that the Finnish study was not designed to look at either autism or inflammatory bowel disease. He confirmed that the study was not specifically designed to look for autism, as no-one had ever raised this issue at the time.
Ÿ The Peltola study simply identified the 173 children (out of 1.8m persons, including troops), who had acute reactions to MMR, then followed only these children up. The study followed up the wrong children. No-one has ever suggested that autism follows an acute reaction.
Ÿ There would almost certainly have been potential autism cases amongst the remainder of the 1.8m, but these were missed, because they were excluded from the study, as it had a 3-week cut-off for reporting reactions. After that point, the remaining (theoretically, 1,799,827) children/other persons were ignored.
Ÿ Peltola relied on referrals from health workers out in the field, who would never have connected degeneration into autism, several months/years after MMR, as being a potential adverse reaction to a vaccine. The alleged syndrome was not known of by scientists, let alone by health-workers in the field, at that time.
Ÿ The UK DoH interpretation of this study, widely trumpeted during 1/2001, is that Peltola "clears" MMR of a link with autism/IBD. It is difficult to accept that this "conclusion" has any degree of scientific justification. It appears that the DoH's "conclusions" have been retrospectively bolted-onto an old and irrelevant study.
There are other awkward facts regarding the Peltola study:
Ÿ The study was part-funded by Merck Sharp Dohme (MMR manufacturers).
Ÿ The study barely refers to autism or IBD.
Ÿ Reviews of the study (eg December 2000 Medscape) do not even mention autism/IBD, which are obviously not seen by the reviewers as a relevant aspect of this study.
Despite this, the Peltola study continues to be cited by the UK medical establishment as conclusive proof that there is no link between MMR and autism. As late as 12/2001, Dr. Simon Fradd of the General Medical Council's Doctor-Patient Partnership quoted this study by Peltola on BBC Radio 4 as conclusive proof of the absence of any link.
The UK DoH also said in a personal communication, referring to all the various studies: "the follow-up time (three weeks) was based on knowledge of the replication rates of the vaccine viral components.....it is recognised that such a study could not establish a causal relationship with extremely rare events..... millions of children have received MMR in other countries such as Finland and the USA; no serious long-term complications have been identified...." (my emphasis).
Overall verdict: this study is wholly irrelevant to the issue of whether MMR can cause autism.
43. The Kaye, Melero-Montes and Jick Paper, MMR Vaccine and the Incidence of Autism Recorded by General Practitioners: A Time Trend Analysis, British Medical Journal, February 2001
This paper attempted to prove that there was no link between MMR and autism because, although autism increased when MMR was introduced, it has carried on increasing since, even though MMR's coverage reached near-saturation almost immediately after its introduction into the UK in late 1988.
Ÿ The study looked at 305 children (254 boys) aged 12 or under with autism diagnosed in the years 1988-99. It also looked at 114 boys aged 2 to 5 years born in 1988-93. It used the UK General Practice Research Database.
Ÿ The study found that autism had increased sevenfold from 0.3 per 10,000 in 1988 to 2.1 per 10,000 in 1999 (note how low this figure is compared with other studies)
Ÿ In the 114 boys born 1988-93, it found autism had increased fourfold, from 8 per 10,000 (1 in 1250) for boys born in 1988, to 29 per 10,000 (1 in 345) for boys born in 1993, during a period when MMR take-up was claimed to be constant at around 97%.
Ÿ The study concluded that no correlation existed between MMR and autism, and that the explanation for increased autism remained uncertain
Ÿ However, the authors acknowledge that their methods were a "second-best", because what they really wanted to do was compare vaccinated and unvaccinated cohorts of children. They said that this was impossible because only 3% of cases and controls did not receive MMR. Given the very small numbers of autism cases they in the event actually looked at, this seems an unconvincing argument for abandoning their preferred approach
Ÿ The authors then argue that if MMR was a major cause, then the risk of autism should have stopped rising within a few years.
Ÿ However, they also admit that the diagnosis of autism was not confirmed from original records, but conclude that "differential misclassification of the diagnosis in vaccinated and unvaccinated children is unlikely to vary over the period of the study", though no evidence is offered to back this claim.
Ÿ They also acknowledge that time trend analysis is a "relatively crude method".
Ÿ The study authors go on to speculate that the increase in autism that they found "could be due to increased awareness of the condition among parents and GPs, changing diagnostic criteria or environmental factors", without subjecting these "explanations" to any detailed scrutiny.
Ÿ The authors also acknowledge the further limitation that they have not yet obtained and evaluated full clinical record information from GPs to describe more fully the characteristics of children diagnosed with autism and to explore other possible explanations. Yet they still dismiss MMR, despite this shortcoming.
Ÿ It might be the case that the increase in autism that the authors find, over the period 1988 to 1997 (note - not 1999 - the study figures actually fall away after 1997) could be due to a hybrid explanation, with increases in the early years due to MMR and then continuing further increases in the later years due to better awareness. There is nothing in the study to refute this criticism
Ÿ It is also unclear how the issue of re-vaccination has been dealt with. What of the seven million children vaccinated or re-vaccinated in 1994 in the UK "Operation Catch-Up" programme? Couldn't the continuing rise in diagnosed cases in 1995-97 be due to Operation Catch-Up? The study does not mention it.
Ÿ It is interesting that the Finland study team (Patja et al) said "Causality between immunisation and a subsequent untoward event cannot be estimated solely on the basis of a temporal relation." Yet the Kaye et al study uses a basically similar approach to "prove" there is no link, comparing temporally-linked trends in MMR take-up and autism increases.
Ÿ There is also a question over the use of mercury-based preservative (thiomersal, or thimerosal) in vaccines. This was used in the late 1980s and early 1990s, but has reported to have been largely phased-out in the US, with a free exchange system being operated by the manufacturers. No such exchange has operated in the UK, with existing thiomersal-based stocks being used up on the children. Autistic enterocolitis may involve thimerosal as part of the damage sequence.
Ÿ If it did, and following a change in formulation, then this might well explain continued rises in autism through the 1990s, then a fall-away in increases at the end of the decade, as was actually found by Kaye et al. Did the industry change the preservative formulation as public concern grew? And has this affected the statistics of autism?
Overall verdict: this study offers no convincing evidence against an MMR/autism link.
44. Paper by Dales, Hammer and Smith, Time Trends In Autism and in MMR Immunisation Coverage in California, Journal of the American Medical Association, March 7th 2001 Vol. 285, No. 9, 1183-1185
This paper, entitled "Time Trends in Autism and in MMR Immunisation Coverage in California" is one of the least conclusive and least robust of all the research of recent years. It appeared in JAMA, March 7th 2001, but it is surprising that it achieved such a high profile within the UK, so weak was its hypothesis and so inconclusive its contents.
The paper attempted to determine if a correlation existed in trends of MMR immunisation coverage and autism occurrence. It did this by examining data from 21 regional centres covering the whole of the State of California.
During the years examined, 1980-94, MMR take-up was about 72% prior to 1988 and about 82% after 1988. Autism increased from about 200 in 1980 to about 1200 in 1994. The trend of increasing autism continued after the introduction of MMR and was claimed to be unaffected by the increase in take-up.
This hypothesis, of a correlation, could be criticised as not being useful to the detection of any MMR/autism link. Although immunisation coverage can be determined, with a specific "date of immunisation", autistic spectrum disorder ranges from the mild to the severe, its onset ranges from the rapid to the gradual, and its diagnosis varies from a timely and accurate diagnosis to no diagnosis whatever. This apparently was not taken adequate account of by Dales et al.
The study did acknowledge some weaknesses itself:
Ÿ "Diagnosis is not always straightforward". This is an extreme understatement.
Ÿ "California Department of Developmental Services' report stresses that its patient caseload data cannot be used as a true measure of changes over time in autism incidence because other factors can affect trends in system case numbers"
Ÿ "Observation of parallel trends over time.......generally do not constitute strong evidence for a causal association between the two events"
Ÿ "As the system expanded and matured over time, the proportions of California children enrolling and the distribution of ages at enrolment likely (my emphasis) changed over time as a result". Clearly, the authors do not know, one way or the other, not do they attempt to quantify this to enable their reliance on the data to be validated, or appropriate potential distortions in the data eliminated.
Ÿ "Also, the proportions of children enrolling in the system who were born outside California may (again, my emphasis) have changed over this time period". Again, they do not know, have not attempted to quantify this factor, and cannot correct for it.
Ÿ "The data presented herein have some limitations. It would have been useful to examine individual immunisation and autism records on the same children; however, these could not be linked". What the authors are saying here is, they would like to have done a rigorous study, but they couldn't obtain the data.
Ÿ "Further, the childhood immunisation coverage data used in this study do not provide precise quantification of the percentage of children who received the combined MMR vaccine product vs. separate injections". This is an admission that one element of the two elements that provide the statistical comparison that is central to their hypothesis, is inaccurate. They go on to say that historical data from elsewhere in the US "strongly suggests" that the use of separate vaccines was "rare" for the 1984-94 birth cohort. How strong? How rare?
Ÿ Despite this catalogue of drawbacks and "softness" - or complete absence - of data, the authors then go on to claim that they have been "unable to demonstrate a correlation between secular trends in early childhood MMR immunisation coverage and autism caseload". A dispassionate and objective observer would find this wholly unsurprising.
Ÿ The assumption that there would be a plateau in the increase in MMR (to match a plateau in take-up of MMR) would only be valid if the background susceptibility of the infant population has remained constant. If successive generations of children became increasingly susceptible to an adverse event such as autism, caused by MMR, then this might well be reflected in a continuing rise in autism. This obvious possibility is not addressed. It does not have to be the case that the relationship between MMR and autism is a simple linear one, without other factors being involved.
Overall verdict: this study is not relevant to disproving an MMR/autism link If the study does have a value, it is to demonstrate that extremely weak studies are not only capable of achieving publication - apparently without attracting peer-review criticism - but also that they are then uncritically welcomed, and publicised, by one side of the argument. This in itself is illuminating.
45. Paper by DeWilde, Carey, Richards et al, Do Children Who Become Autistic Consult More Often After MMR Vaccination, British Journal of General Practice, March 2001
This paper appeared in the British Journal of General Practice, March 2001. It attempted to test the hypothesis that a degeneration into autism, with subsequent diagnosis, would be reflected in increased consultations with the child's general practitioner.
This would appear to be an extremely weak hypothesis to test. For example:
Ÿ It may be difficult to place a definite date upon degeneration
Ÿ Parents may not seek assistance from their GP immediately, or even at all in some cases
Ÿ Parents may seek advice from health visitors or other health professionals
Ÿ Parents may see a GP only once, to obtain a referral to a specialist paediatrician
Ÿ Parents may see their GP for reasons unconnected with autism, confusing the data in some cases
Ÿ Parent may be extremely reluctant to see their GP, because of the sometimes extreme practical difficulties of taking an autistic child to a public surgery, with waits etc.
The study authors do not acknowledge any of these serious potential methodological flaws, nor do they attempt to quantify them in an attempt to validate the effectiveness of their methodology.
The authors looked at only 71 cases of autism, a small sample by any standard for testing a statistical hypothesis, and identified numbers of consultations from a primary health care database. It found that there was no significant difference between cases and controls for numbers of consultations in either the six months before/after immunisation, or the two months before/after immunisation.
The study also noted
Ÿ that there was a significant fall-off in consultations in the six months after immunisation, in both cases and controls. However, it did not address the possibility that this might have been for two entirely different reasons, with healthy children not needing to be taken to their GP, and autistic children not being seen by their GP for other reasons such as those set out earlier. The study simply assumed that the fall-off in the cases and the control group was for the same reason, without evidence to underpin this assumption.
Ÿ It acknowledged that it could be criticised because the study authors "cannot confirm that our cases truly suffer from autism"
Ÿ The study, like almost all other studies that "prove" no MMR/autism link, did not specifically address the cohort of children alleged to have degenerated as a consequence of MMR, and who are now proceeding through the legal processes
Ÿ It acknowledged that "some diagnoses will have been missed"
Ÿ It admitted that "it seems unlikely (my emphasis) that these will be specifically those associated with MMR", although it offers no evidence to support this assumption.
Ÿ The study notes that ""the clear difference in consultations in the six months before the diagnosis of autism" (between cases and controls) "emphasises that consultations were being recorded and that differences in consultation rates between cases and controls were detectable". But the study does not address the possibility that the higher frequency of consultations by cases is linked to a potentially-associated condition, such as otitis media (and consequent antibiotic use), and that cases moved from more frequent consultations than controls for such a condition, to more frequent consultations than controls for a wholly different and more serious condition.
Overall verdict: this study is not relevant to disproving an MMR/autism link. In short, there are so many caveats, acknowledged and unacknowledged shortcomings and other methodological limitations to this study that its conclusions are virtually valueless. Again, it is illuminating that it has been so well received by one side of the debate (the UK Department of Health).
46. Study by Davis et al, Measles-Mumps-Rubella and Other Measles-Containing Vaccines Do Not Increase the Risk of Inflammatory Bowel Disease, Archives of Pediatrics and Adolescent Medicine, 2001, 155: 354-359
This study was conducted in the US on the populations of four health maintenance organisations as part of a vaccine safety programme co-ordinated by the Centres for Disease Control and Prevention.
The study focussed on the following questions:
Ÿ Was the age of first vaccination with MMR or other measles-containing vaccine, or receipt of vaccination itself, associated with an increased risk for Crohn's disease or ulcerative colitis later in life?
Ÿ Was receipt of MMR or other MCV associated with the acute onset of disease shortly following vaccination?
In each of the areas, trained staff reviewed medical records. Cases were of individuals enrolled since birth (some as early as 1958) to 1989. It was claimed that consistent criteria were used for definite and probable diagnosis of Crohn's disease, ulcerative colitis or unspecified irritable bowel disease. This involved diagnosis by a gastroenterologist, "with signs and symptoms and a diagnostic test for IBD". Five controls were selected for each case, matched by sex, health organisation and year of birth. Dates of vaccination, type of vaccine and date of diagnosis were also recorded.
There were 155 cases of IBD with 152 definite or probable cases. Seven had no discernible onset, two were of "unspecified disease" and one was vaccinated when older than 10 years. This left 142 cases and 432 controls for further analysis.
The study found that:
Ÿ the risk of inflammatory bowel disease was the same whether for vaccinated or unvaccinated people
Ÿ there was an average of 140 months between vaccination and diagnosis for cases.
Ÿ Only 1% of cases developed inflammatory bowel disease within a year of vaccination
Ÿ Only 1% of controls developed inflammatory bowel disease within a year of vaccination.
Ÿ Whether children were vaccinated before 12 months, between 12 and 18 months, or after 18 months, showed no difference in the risk of developing inflammatory bowel disease
However, the study team had to acknowledge several serious limitations to this study:
Ÿ Only patients with a physician diagnosis (usually a gastroenterologist) were included. This could have potentially missed many cases, particularly if those missed were of an insidious new variant
Ÿ The team acknowledged the inherent limitations of diagnostic accuracy in any retrospective study
Ÿ They had little information on children or adults with non-specific colitis that did not lead to an eventual diagnosis of IBD - surely a key failure, given the nature of the research by the Wakefield team at the Royal Free Hospital in London
Ÿ There was an acknowledged limitation over statistical power. The report admitted:: "We were able to effectively rule out associations larger than 2-fold between ever being vaccinated with MMR and developing IBD, and associations larger than 3-fold between vaccination with other measles-containing vaccines and IBD. However, we had a limited sample size from which to look at the independent associations between vaccination and either Crohn's disease or ulcerative colitis, or at the relationship between timing of vaccination early in life and subsequent risk for Crohn's disease or UC." This seems to be a serious self-criticism, yet oddly it does not seem to have had much effect on the study's assertive conclusions.
The study's reliance on patient records should also be questioned. The analysis of records can by definition be only as good as those records themselves. No study (as far as is known) has yet endeavoured to verify whether children suffering from acquired autism, ileal lymphoid nodular hyperplasia or non-specific colitis have medical records that accurately reflect these conditions. There are grounds for suspecting that the very reverse may be the case. The difficulties in obtaining a clear and timely diagnosis of autism are well known. The nature of the autism problem, with many patients without speech, means that the precise nature of the patient's complaints and symptoms may be poorly recognised, and even more poorly recorded.
Overall verdict: although this study at first sight appears more persuasive than some others, it too fails to provide convincing evidence against an MMR/autism link. The study may be seriously flawed due to its retrospective nature, when the condition in question (acquired autism after MMR/MCV) has only recently received publicity, and because of doubt over records.
47. Further Paper by Farrington, Miller and Taylor, MMR and Autism: Further Evidence Against a Causal Association, Vaccine, 19 (2001) 3632- 3635
When it became apparent to Taylor, Miller and Farrington that the time-lapse for degeneration into autism might be a protracted one, they were obliged to re-analyse their earlier data.
Ÿ Farrington, Miller et al repeated their view of the original Wakefield study, that it was very small (12 children) and that the interval between receipt of MMR and first behavioural symptoms varied from 24 hours to two months. However, the Wakefield study cohort subsequently grew to about 200, and this has not been acknowledged by Farrington, Miller et al in this further paper.
Ÿ The Farrington, Miller et al study also has not taken account of the Spitzer, Aitken et al study and its implications (see later sections). They also maintain that they "found no evidence to support a causal association". But they themselves, in their first study, unconvincingly dismissed a clustering of parental concerns at around six months. They maintain this unconvincing stance.
Ÿ Farrington et al concluded that the temporal association found by Wakefield et al was "a combination of selection bias and chance". This latter is a highly contentious conclusion, suggestive of wishful thinking, in the same way as the dismissal of the six-month clustering was.
Ÿ In this second paper, the authors seek to re-test their earlier conclusions by removing any preconceived fixed-time interval between vaccination and the onset of autism. Again, they use a statistical methodology, self-matched case-series analysis, but once again with a very small (for this method) data set of just 64 cases of what they describe as "unvaccinated" children with autism - presumably, they mean "unvaccinated with MMR" - plus 231 cases of children with autism who had received one dose of MMR, and a further 62 cases of children who had received two doses of MMR (total 357 children).
The study found that:
Ÿ for the 357 cases, the observation periods had a median of 89 months, a maximum of 191 months.
Ÿ The oldest age at diagnosis was 180 months.
Ÿ Some 64 did not receive MMR.
Ÿ Some 43 received MMR after age 2 years, at median age 57 months, maximum 165 months.
Ÿ Some 62 cases received a second dose of MMR, at median age 54 months, maximum 159 months.
The comparison of relative incidence for each group finds that there is little difference between those that had received MMR and those previously referred to as "unvaccinated", but which seems to have really meant "vaccinated with single-antigen measles vaccine" - the paper is not clear.
The major criticism of the earlier paper using this data (see above section) were that there was only a proxy for "onset of autism" (a questionable term in itself). The original study measured diagnosis, parental concern and regression (if applicable) from medical records. But these would be variably delayed from any actual "onset event". The very poor correlation between these proxies and the "event" means that the analysis loses all statistical power.
Major criticisms of this further re-worked paper's statistical methodology are that:
Ÿ Regarding the whole period following MMR as being "at risk" is questionable.
Ÿ Looking to see if those who have MMR earlier have a proxy variable earlier is erroneous, when one observes the very narrow timescale for the application of MMR in this paper. When the input signal (the age of receipt of MMR) has very little variability, one would be unlikely to find this reflected in the output signal (date of diagnosis)
Ÿ The above flaw means that the only statistical power left is coming from finding any difference between those who have MMR and those who have not. But most of those who do not have MMR are those older children who are of the pre-MMR generation. So Farrington et al's analysis is effectively asking whether those who are older had had an earlier or later onset of autism (as measured by the proxy variables).
Overall verdict: this study fails to provide any convincing evidence against an MMR/autism link.
(Note: this study has been claimed by the UK Medical Research Council to represent "strong positive evidence" of there being no MMR/autism link)
48. Paper by Fombonne & Chakrabarti, No Evidence for a New Variant of Measles- Mumps-Rubella-Induced Autism, Pediatrics, Vol. 108 No. 4 October 2001
This paper examined whether there is a new phenotype of autism involving regression and gastrointestinal symptoms.
It is suggested that where this paper is flawed is in the assumptions underpinning the hypotheses that are tested. All else stems from that. Fombonne & Chakrabarti assume that if autistic enterocolitis existed, then one or more of the following six predictions should be supported by empirical data:
Ÿ Prediction (1) - "childhood disintegrative disorder has become more frequent". (The study found the prevalence of childhood disintegrative disorder to be 0.6/10,000, or 1 in 16,666. But this seems far too low in comparison with other recent studies).
Comment - historic data is not available to prove this either way. The claim that the present rate of 1 in 16,666 represents no increase is further undermined by its non-credible low level. Other studies have found rates very many times higher. This strongly suggests that the study is flawed.
Ÿ Prediction (2) - "the mean age of first parental concern for autistic children who are exposed to MMR is closer to the mean immunisation age than in children who are not exposed to MMR."