Monday, 30 July 2007

What happens to autism rates if we stop MMR?

From Honda et al 2005 J Child Psychol Psychiatry 46(6):572-9 (without permission):

"The Japanese MMR vaccination program targeted one-year-olds between April 1989 and April 1993, then was discontinued. Therefore, children born during the years 1988 to 1992 received the MMR vaccine in years 1989 to 1993 at one year of age. According to Yokohama statistics, MMR vaccination rates declined from 69.8% in the 1988 birth cohort, to 42.9%, 33.6%, 24.0%, and a mere 1.8% in birth cohorts 1989 to 1992...the seven-year cumulative incidence of ASD rose progressively from 47.6 per 10,000 for children born in 1988 to 117.2 for those born in 1996, that this rise continued in cohorts of children born after MMR was withdrawn, and that no decline in ASD incidence occurred in the five-year period from 1988 to 1992 during which MMR vaccine usage fell from 69.8% to zero population coverage...Accordingly, it is possible to conclude that it is extremely unlikely that MMR has been responsible for the rise over time in the incidence of diagnosed autism. It follows that it is similarly unlikely that it causes autism frequently or at all. It cannot have caused autism in the many children with ASD in Japan who were born and grew up in the era when MMR was not available. Because this frequency is at least as high as in populations in other countries in which most children were vaccinated, it implies that MMR could not cause a substantial proportion of cases of autism."

That is, when Japan stopped MMR its rates of autism kept on rising, which is not what we would expect if the rise in autism rates (if they are real rises and not due to changing diagnostic patterns, see below) were due to MMR.

Sunday, 29 July 2007

Autism and "mental retardation"

I've plotted the autism prevalence against the "mental retardation" rate from the Croen et al study mentioned below (numbers just estimated from the figure). The correlation coefficient is -.5 and is not significant by conventional criteria.

Changing diagnosis of autism

In pursuance of an argument on badscience, I'm reproducing (without permission) a figure from Croen et al 2002 J Autism Dev Disord 32 (3):207-15 showing the temporal relationship between diagnoses of autism and mental retardation over time. This relationship suggests that changing categorisation of behaviour may underly some of the reported increase in autism prevalence.

Friday, 27 July 2007

Ethics of animal research

I think the view of the British public is that raising, killing, and eating animals for food and other animal products is perfectly normal, natural even, and nothing to be concerned about. On the other hand, animal research, or vivisection, is a rather distasteful furtive practice that is only tolerated because of the benefits it brings humanity - but must be constantly monitored and regulated to keep it in check.

They're only half right. Their concern for fluffy rabbits and cute monkeys doesn't seem to extend to battery hens or other food animals and this is a rank hypocrisy motivated entirely by a refusal to think about the issues coupled with an irrational sentimentalism. I'm with Peter Singer that animal welfare has to be part of a utilitarian ethic, and based on current farming practices eating meat and animal products fails that cost-benefit analysis.

So why am I in favour of animal research? The reason is pretty simple, on a cost-benefit calculation animal experiments can bring benefits to humans, organisms that weigh very heavily in my moral calculus, at the cost of animals, which weigh somewhat less (depending on species and experiment performed), and these benefits cannot be got through other methods. Contrast this with eating meat which is pretty much unnecessary for me in my rich Western environment, and where rearing and slaughtering methods are far from halcyon.

Most people would probably accept my analysis of the need for animal research, even if they reject my arguments for vegetarian/veganism. And this is the fundamental problem of the anti-vivisection movement, they can put forward a logically consistent, even respectable argument against animal research, often based on a position that essentially rejects privileging humans over other animals (although not necessarily), but most people will never accept their arguments.

So what are they to do? Fortunately for them, there is another line of attack. If the animal suffering end of the equation isn't powerful enough, how about questioning the benefit to humans? Now this isn't entirely an opportunistic strategy, many anti-vivisectionists have convinced themselves of the futility of animal research for benefiting human health because of the cognitive dissonance inherent in the (intellectually respectable) position that whilst animal research may benefit human health, these benefits are outweighed by the suffering of the animals.

Many anti-vivisectionists go further than just asserting that animal research is useless, they want to argue that it is actively harmful to people's health. An additional line of attack is to assert that we don't even need to do animal research, there are in fact a range of super-dooper alternatives that are being ignored by the pharmaceutical companies and evil scientists with their vested interests (these include 'computer models' and 'in vitro' tests, it is never entirely clear how these are supposed to be developed, and how exactly they are meant to work, but the anti-vivisectionists seem quite taken by these mythical beasts).

Now I've spent some time working on researching disease in humans, yet I have, on occasion, carried out animal research in pursuit of this, and constantly rely on the results of studies into animals when interpreting, planning, and evaluating my own work. In fact, if you ask doctors and medical researchers the vast majority will stress the value and relevance of animal studies. Without animal research I wouldn't even know where to begin studying the human brain - it would just be a big squidgy blob of tissue.

Now anti-vivisectionists are not entirely wrong that a whole lot of research goes on that is of, shall we say, questionable relevance to human health (and I've argued before that science could do with a little more direction, in terms of funding). And I can accept that perhaps the existing regulations are not as well enforced as they could be (although I do think the legislative framework is pretty good, and justifiably strict - if unnecessarily bureaucratic). But they go way beyond this. I have a familiarity with the pseudo-academic works of Shanks & LaFollette and Ray & Jean Greek, as well as the continuum from national campaigning to street level demagoguery of the BUAV, SPEAK, Animal Aid, X-CAPE and a host of others.

I feel like I should present some critique here of their overall position, but it is really impossible because they are generally so incoherent, scattergun, and ill-informed that they can only be countered on an argument by argument basis. Not that I would advise actually trying to engage any of the true believers as they are impressively inflexible in their thinking and quite unable to connect with a rational argument. However, there is a loose thread that runs through these arguments, and that is the view that animals are so different from humans that we simply can't infer anything about how humans work from looking at animals - now stop snickering at the back and muttering 'DNA' into your hand these people are serious - they draw a lot of their intellectual strength from Shanks & LaFollette and the Greeks, but the argument is essentially that because humans and animals are not exactly the same you can't extrapolate from one to the other. That is pretty much it, with a variety of anecdotes about how this or that drug works differently in humans than this or that species, oh, and mentioning thalidomide of course (yes, I know that thalidomide was never tested for teratogenicity in animals - but they apparently don't).

To get an idea of the level of argument we're talking about here's an exerpt from a BUAV 'report' on animal research in Parkinson's disease, and compare their claims with the development of deep brain stimulation for Parkinson's disease using animal models:

"From a scientific point of view, animal 'models' of PD should be stable over a long period of time so that the effects of new therapies can be evaluated. However, the fact is that the signs and symptoms in primates vary over time, even between individuals. Additionally, specific signs of illness in individual animals used as indicators of the severity of their condition contradict each other, making data interpretation more difficult. There are serious underlying limitations to the primate 'models' of PD which include:

  • Symptoms are caused by toxin injection and appear rapidly in primate 'models'. The causes of human PD are unknown and symptoms are
    slow to develop.
  • Because of the artificial causation of the condition in monkeys, little can be learned of the causes and progression of the human disease.
  • Compensatory mechanisms in surviving brain regions are likely to be different in lesioned monkeys compared to PD patients.
  • If dosing is stopped, neurotoxin-treated primates show partial but variable recovery (and there is variation between old and new world monkeys). However, humans always show a progressive worsening of the symptoms over time.
  • PD usually affects older people with comorbidities, while the monkeys used in research are young and otherwise healthy.
  • In the affected brain regions of primates, Lewy bodies are either never seen or only very infrequently. Yet in patients, these cellular inclusions are the classic hallmark of PD.
  • MPTP-treated primates show limb tremor only sporadically. In PD patients tremor is marked and sustained. The cognitive patterns of impairment also differ between primates and patients.
  • In neurotoxin-treated monkeys, specific dopamine-containing brain cells are damaged in one part of the brain. In PD, damage is more widespread and involves other neurotransmitters, in addition to dopamine.

Replacing primates in fundamental PD research
Functional imaging has a key role to play in human studies of Parkinson's disease, thereby avoiding problems of species differences and artificiality of the model. Positron emission tomography (PET) imaging has been used to measure levels of dopaminergic activity in the brains of Parkinson's patients. This sheds light on the pathophysiology of the condition, and permits direct study of disease progression at a biochemical level. PET imaging has revealed disturbances of brain functional interactions and cognitive information..."

If you're familiar at all with neuroscientific research, and Parkinson's research in particular you'll be able to see what crap the BUAV are talking moaning about fairly peripheral issues to do with disease progression, whereas the scientists went into the the brains and found a bit that lesioning or stimulating could actually control the symptoms and then translated that into human patients (take a look here at what it can do for real live human patients). No matter how many PET studies you perform you really can't find this kind of stuff out without getting inside that brain - you are not going to look at a PET scan, say 'oh, that bit looks like it lights up a lot in PD, let's cut it out of some people and see if they get better'.

Of course these animal models are not exatly the same as the human disease, but they're not trying to exactly reproduce the disease, they are simply trying to reproduce the relevant problem (dopamine depletion) and find relevant solutions (deep brain stimulation). I could go on about how the normal physiology of the brain can only be discovered by research in animals, and about how much we've learned from animal studies that we've been able to use to interpret the pathology in human disease, but examples of successful research like this are better than any argument I can produce to show how unfounded this sort of stuff from the BUAV and friends really is.

Animal wrongs

The BUAV have got plenty of publicity for their case against the government in the high court. Yet they only seem to have won on one issue:

The High Court ruled that Buav was correct in stating that the government was failing to correctly determine the severity limits for animal experiments.
Which, judging from the BUAV press release refers to

The BUAV is also questioning why the Home Office assigned a ‘moderate’ suffering banding to experiments which included highly invasive procedures such as removing of the top of marmoset’s heads to induce strokes. The guidelines state that any procedure which ‘may lead to a major departure from the animals’ usual state of health and of well-being’ must be categorised as ‘substantial’, and undergo far stricter assessment to get licensed.
Despite this the general tenor of the media coverage is that (e.g.):

The British Union for the Abolition of Vivisection today won its High Court claim that the Government is failing in its legal duty to ensure animal suffering is kept to a minimum in UK laboratories.
Now I'm not sure of the legal minutiae of this case, but my copy of Wolfensohn & Lloyd lists, for moderate severity procedures:

    • Surgical procedures where post-operative care and analgesia are reliably provided
    • Toxicity tests with defined humane end points (as opposed to lethality as the end point)
and for substantial severity:
    • Major surgery causing post-operative suffering
    • Toxicity studies with significant morbidity or death as an end point
    • Any procedure which results in significant deviation from the animals' normal state of health
I think this puts some context to that last phrase, "Any procedure which results in significant deviation from the animals' normal state of health", which, if you weren't aware of the definitions, you might think basically applies to any surgical intervention you might make (it seems like it is likely to cause 'deviation' from it's normal state of health), but clearly that is not the case, otherise the moderate category wouldn't exist.

Of course, the BUAV may well be right that in this particular case the procedure (inducing strokes) did warrant inclusion under the significant category, but my experience with animals receiving neurological insult for experimental purposes suggests that they are generally surprisingly unaffected by the whole business - this is partly because the effects of stroke on an animal are less severe due to the way their brains are wired, and partly due to their general lack of awareness of any deficits induced (not being human and thus unable to ruminate on their loss of capacity).

Thursday, 26 July 2007

Science vs. the media

Francis Sedgemore on the Guardian's Comment is Free site blames the whole MMR debacle on scientists.

Most of his article is content-free rhetoric, but analysing the substance of what he actually has to say:
"in the dock should also go those medics who back in 1998 set out to silence the now discredited Wakefield and his supporters in the most vituperative attack, and fed the media feeding frenzy"
But of course Wakefield was not only wrong, his claims were unfounded, it would have been remiss of them not to have disagreed with him - yet Sedgemore seems to think that they shouldn't have challenged Wakefield's argument, with all the public health ramifications that held. How the hell would that have helped, would the media have not run the scare story? Don't make me laugh. I'm also not sure where this 'vituperative' attack was from, at the time people mostly recommended more research, the attacks on Wakefield personally have come since he has been proven to be wrong, unethical, incompetent, and refused to change his mind when the contrary evidence has come in.

"Medical research does not have a particularly good track record."
Well it hasn't been bad what with those antibiotics and vaccines and all!

"The drug industry routinely engages in disease mongering,"
And this is the fault of doctors and not the complicit media why exactly?

"dietary advice changes on an almost weekly basis,"
Again, does the advice from the medical associations or government change or does the media just spin the findings from each and every study to make it seem like the advice is constantly changing to those unfamiliar with how science progresses slowly based on the weight of accumulated evidence (people like Sedgemore it would seem)?

"bow-tied hepatic specialists advise the government on alcohol taxation,"
So he objects to medical professionals (BMA, chief medical officer) intervening in the public debate on taxing alcohol (an argument to be had, but hardly damning of medical research!) or does he disagree that alcohol is damaging to the liver?

"and dodgy data on cannabis strength are fed to a sensation hungry media in an attempt to have the drug reclassified and users recriminalised."
Is that anything to do with the medical establishment or due to politicians and a complicit media again? As far as I know the medics mostly publish scientific studies on the harmful effects of cannabis, and relatively few offer opinions on legal classification - of those I think the majority are probably against reclassification. But, again, does he disagree with the medical science or just the involvement of medics in the debate about classification?

"But while medics sneer at hacks, they might also like to attend to the planks obscuring their own vision, and work on their serious public outreach problems."
The motes and beams reference is just too funny, all a "public outreach problem", if the media essentially make up their own unjustified interpretation of an unpublished leaked unfinished scientific study and spin it as a scare story and refuse to retract it more than once, that is the fault of the scientists who were not consulted about the story which attributed views to them, and ripped off their research, all whilst making fundamental mistakes in interpreting the data.

I note that Francis Sedgemore tries a much more temperate response that simultaneously fails to respond to the criticism of his hyperbolic attack on medicine and science in defence of journalism. You might almost suspect that his original article was deliberatly over the top in order to provoke people to read and respond, even though it was essentially wrong and unjustified. All a bit redolent of MMR and the media really.

Looking over some previous articles he seems to have a bee in his bonnet about the "medical establishment", particularly regarding cannabis, yet while his articles see him railing against these evil doctors he doesn't seem to refer to anything concrete that they're supposed to have done, e.g. here where he says "Spreading scare stories about cannabis is grossly irresponsible and risks bringing the medical profession into disrepute." but only talks about journalists and politicians.

Elsewehere he harks on a similar point about alcohol taxation:

"...only fools and medical professionals think that the solution there is to raise alcohol prices to Swedish levels, and further restrict the availability of booze. Criminalising cannabis users will likewise not make for a healthier society."
Again, it isn't clear why he disagrees with the argument, or who exactly he is disagreeing with, so I'm not sure where his "Medical research does not have a particularly good track record" argument comes from, or indeed what his point is other than that he doesn't like medics interfering with debates on public policy by giving their opinion, presumably that is exclusively the domain of journalists?


Amusingly Sedgemore links to this thread where I and others argue that the Lancet is partly to blame for the MMR furore and seems to think that supports his position, I'm afraid that won't wash Francis, there is plenty of blame to go around, the Lancet may have helped spark the row but the mainstream media jumped in feet first and is still stoking the fire, which is where we came in. [Sedgemore summarises his response to criticism of his article here]

Wednesday, 18 July 2007

Autism and the gut

Ben Goldacre's dissection of the woeful Observer autism article (defended by the Observer's readers' editor, and oh so unconnected with the same journalist's interview with Wakefield before his GMC hearing) has finally been published in the Guardian and the BMJ.

I don't need to bang on any more on that topic, but something that has baffled me is the continuing support for Wakefield from so many parents. What is it that the MMR 'link' with autism does for them? I suppose it gives them someone to blame (or sue), and perhaps lets them off the hook if they have the misguided worry that it is somehow their fault (say from upbringing or genetics), although I'm not too sure how it is anymore palatable than that they gave their child autism by giving them the MMR.

Of course there are always cranks and crackpots, and the anti-vaccine movement has a long and not so illustrious history preceding the MMR controversy, but looking at the comments from these parents in many of the above discussions it struck me that something many of them mentioned was the gut symptoms that their autistic children showed. This has been something of a focus of Wakefield, who originally set out to blame Crohn's disease on measles infection. So how many people would we expect to show this association?

MMR is given in two injections at ages 1-5yrs, on a back of the envelope calculation (assuming no connection, which may well not be valid), if we just look at proper inflammatory bowel disease rather than just vague gut symptoms we find that taking annual incidence of 6/100k for IBD, and 8/10k for autism spectrum, and estimating around 4 million of the little blighters in the UK, that we'd only expect about one child to show both over that period.

However, if we broaden our coverage to gut symptoms in general we get a rather different story, it turns out that GI symptoms are rather common, with some 16% of 10-11yr olds (you find the data for the under 5s!) fulfilling criteria for recurrent abdominal pain, and 2% showing lymphonodular hyperplasia on endoscopy, suggesting several hundred autistic spectrum children with lymphoid changes in the 1-5yr old age range. Funnily enough it is ileal lymphonodular hyperplasia that Wakefield reports as being associated with autism (with gut symptoms), rather than true IBD, and it looks like he will have a steady supply of patients with the right symptoms and timing to keep supporting him.

An interesting review appeared recently in Histopathology:

Aims: To review the literature on the histopathological diagnosis of the condition termed ‘autistic enterocolitis’.

Methods and results: We have reviewed all published works where mucosal biopsy specimens from autistic children have been examined histopathologically. Abstracts were excluded. Our review of the published works, nearly all from a single centre, identifies major inconsistencies between studies, lack of appropriate controls and misinterpretation of normal findings as pathology. Ileal lymphoid hyperplasia may be more prevalent in children with regressive autism but is also seen in children with food allergies and severe constipation, the latter being an extremely common finding in autistic children.

Conclusion: The histopathological diagnosis of autistic enterocolitis should be treated with caution until a proper study with appropriate methodology and controls is undertaken.

And concluded:
There is no doubt that autistic children suffer considerable gut symptoms and these have a significant effect on their quality of life. But is the claim that these children have an underlying IBD justified? We can only speculate why severe constipation was not acknowledged as a significant gut symptom in the original paper1 and was revealed only in correspondence.8 Having accepted, however, that this was a major problem in these children, subsequent studies from this group should have explicitly included developmentally normal children with severe constipation as the appropriate control group. The conclusions of the 1998 and 2000 papers from the Royal Free must therefore be regarded as unreliable because of the use of inappropriate controls. By the time this group did include a control group with ILNH and constipation and indeed reported the lack of significant difference in mucosal pathology between autistic children and controls,3 media and public opinion were already entrenched.

Significant bias was introduced to the studies from the Royal Free by the interpretation of histological changes seen in normal lymphoid follicles as pathology...It is also disturbing that the investigators have not attempted to re-inforce the histopathological diagnosis of enterocolitis by having the slides examined in an open forum by independent pathologists, especially since another small series has found no abnormalities...

There is no evidence from the papers discussed above that a significant number of autistic children have ‘enteritis’, i.e. inflammation of the small intestine. The only consistent abnormality seen in these children may be ILNH, but we have explained why this finding is not unexpected in constipated individuals...

Evidence has been presented to suggest that autistic children have a ‘colitis’. In their efforts to present convincing findings, however, the authors have failed to use appropriate controls and rigorous methodology, leading to serious flaws and unreliable conclusions...As to the aetiology of the colitis in autistic children, there is no convincing evidence that the changes are due to the autism per se. Indeed, as pointed out in the Royal Free papers, many autistic children have severe constipation and/or food allergy, either of which could be responsible for the pathological abnormalities seen. Other possible causes of inflammation in these children include parasitic infestation and swallowing foreign bodies.


In conclusion, we are highly sceptical that ‘autistic enterocolitis’ is a genuine histopathological entity in children with regressive autism. Further studies with rigorous methodology and appropriate control groups could be carried out, but if these are not possible, the histology slides from the autistic children seen at the Royal Free Hospital should be examined by independent experts in an open forum. The results of this exercise would, in our opinion, show that ‘autistic enterocolitis’ does not exist.