I’ve been looking over Patrick Holford‘s recent e-mail on GETTING TO THE GUTS OF TRUTH ABOUT AUTISM, ALLERGY AND MMR. I’ve been trying to analyse Holford’s evidence base – but, given the shoddy references provided as ‘evidence’, this has been a rather frustrating experience.
Holford seems pretty keen on a gluten- and casein-free diet for people on the autistic spectrum: he approvingly quotes Robert Cade arguing that, when they strictly follow such a diet in order to reduce the levels of peptides in their blood, “most patients either improve dramatically or become completely normal”. I’ll put aside, for a moment, the assumption that ‘normal’ is either an unproblematic concept or something to aspire to: there is, at any rate, not convincing evidence that dietary changes can make people on the autistic spectrum ‘normal’.
Holford’s reference for this statement is http://www.panix.com/paleodiet/autism/cadelet.txt. Sadly, however, this link doesn’t work. The closest thing I could find was a link from paleodiet.com to an archived version of Cade’s site (I’m not sure why Cade appears to have taken the site itself offline). There is a tantalisingly named ‘research’ section on this archived site; however, this is disappointing.
The site argues that “for an autistic individual, it has been found that a defect in the intestinal wall permits incompletely digested components of the original proteins to pass from the intestine into the bloodstream”. This could be very interesting – except this is simply asserted, without any evidence being provided.
The site also asserts that “[b]y removing sources of gluten and casein from the diet of autistic children, we have had immense success in at least alleviating and at times eliminating the symptoms of autism.” On seeing this exciting information, I looked over the site for details of research methodology used, control group, randomisation, verification of the elimination of autism, etc – but, oddly, these details seem to have been omitted.
The only other ‘research’ that the site provides us with is a cursory description of a “blood test called a gluten/casein screen”. While the site asserts that “we feel that it provides valuable information on the potential benefits of the diet as well as an opportunity to monitor changes”, it does not provide any convincing evidence that this test does anything useful (or even much detail on how the test might work).
The really odd thing here is that there is some evidence that gluten- and casein-free diets may benefit some people on the autistic spectrum (although not provide any kind of miracle cure, if one wants to view autism as something to cure). I’ve spent way too long chasing up Holford’s shoddy referencing of Cade’s work – but a couple of minutes of searching pubmed led me to the excellent 2004 Cochrane review of the data. When analysing the research on such diets, the review found that only “one trial met the criteria for inclusion” (none of Holford’s ‘evidence’ for the benefits of this diet met the criteria for inclusion in the Cochrane review). The review concludes that there should be more research in this area but there is some – small scale, limited – evidence that such diets may bring some benefits:
Extensive literature searches identified only one randomised control trial of gluten and/or casein free diet as an intervention to improve behaviour, cognitive and social functioning in individuals with autism. The trial was small scale, with only 10 participants in the treatment group and 10 participants in the control group. Results indicate that a combined gluten and casein free diet may reduce some autistic traits. This is an important area of investigation and large scale, good quality randomised control trials are needed.
Even where there is some evidence that dietary changes may bring benefits for people on the autistic spectrum – and where there is the need for more good quality research in this area – Holford instead prefers to reference older papers and broken links. I’m not at all sure why. At any rate, if this is Holford’s ‘normal’ standard of evidence and referencing, I really don’t think that ‘normality’ is something to aim for.