Paul Shattock & Paul Whiteley
Autism Research Unit, University of Sunderland, UK
These notes do not constitute a recommendation or endorsement of a dietary method to alleviate the symptoms of autism or related pervasive developmental disorder. Any decision to undertake such a method must lie solely with the person with autism or with those having responsibility for their care. It is strongly recommended that anyone considering such interventions seek the support of their Medical Practitioner and, if possible, a knowledgeable dietician or nutritionist.
In the early 1980s a number of researchers, including Herman and Panksepp, noted the similarities between the behavioural effects on animals of opioids, such as morphine, and the symptoms of autism. In a very speculative paper, Panksepp proposed a mechanism whereby people with autism may have elevated levels of opioids which occur naturally in the Central Nervous System (CNS) of humans. The best known of these naturally occurring opioid compounds is beta-endorphin (endogenous morphine) and certainly there is a degree of correlation between the known effects of this compound and the symptoms of autism. Just after this, Gillberg produced evidence of elevated levels of "endorphin-like substances" in the cerebrospinal fluid of some people with autism. In particular, elevated levels appeared in those children who appeared to feel pain less and who exhibited self-injurious behaviour. At about the same time, Reichelt produced evidence of abnormal peptides in the urine of people with autism. We ourselves, like a number of other groups, attempted to replicate his findings. Although his method was comparatively simple there were technical difficulties and these attempts were, initially unsuccessful. Later on we switched to a more sophisticated technique (HPLC) and have been able to provide partial confirmation of Reichelt's findings. In the urine of a proportion of people with autism and related conditions, there appear to be elevated levels of substances with physico-chemical properties similar to those expected from opioid peptides. The quantities of these compounds, as found in the urine, are hypothesised to be much too large to be of CNS origin. The quantities are such that they can only have been derived from the incomplete breakdown of certain foods.
Proteins consist of long chains of units known as amino-acids. Normally proteins are digested by enzymes in the intestines being broken down into these units. However, if for some reason, this digestion is incomplete, short chains of these amino-acids (known as peptides) will result. It is proposed that these peptides may be biologically active and could result in some symptoms similar to which we see in autism. The majority of these peptides will be dumped in the urine, which is where Reichelt and we speculate that we are finding them. A small proportion will cross into the brain and interfere with transmission in such a way that normal activity is altered or disrupted. It may be that these compounds, themselves, have a direct effect upon transmission or that they will attach themselves to the enzymes which would break down our own naturally occurring enzymes. The consequences would be the same in either case. It is well known that casein (from human or cow milk) will break down in the stomach to produce a peptide known as casomorphin which, as the name implies, will have opioid activities. Similar effects are noted with gluten from wheat and some other cereals in which case the compounds formed are gluteomorphins. If this opioid excess hypothesis is correct, there are a number of strategies which can be adopted. Firstly the anti-opioid drug Naltrexone could be considered, and promising initial results have been reported. Not all of the reported trials on Naltrexone have produced positive benefits but where appropriate very low dose therapies are employed the results seem to be better. Alternatively, a diet which excludes casein (milk and dairy produce) or gluten (wheat and some other cereal products) could be considered. It may be possible to determine, from the pattern of the urinary peptides whether casein or wheat or both should be avoided but such conclusions may be premature at this stage. It has been observed that those children whose autism appears at or around the time of birth may have a problem with casein whereas those whose autism becomes apparent at about two years of age, when a wheat based diet is more likely to be adopted, have particular difficulties with gluten. Some children may have difficulty with both. Norwegian colleagues of Reichelt have published data which support the effectiveness of such dietary programmes but these studies cannot be considered as conclusive. Numerous people have experimented on an individual basis and have reported successful responses but such evidence cannot be considered as, in any way, conclusive. During preliminary studies of parental reports, however, the results appear to be very much superior to those obtained with equivocal drug-based therapy.
The theoretical processes described here are toxicological in nature rather than allergic. The results are akin to poisoning rather than an extreme sensitivity such as occurs in coeliac disease or sensitivity to certain food colourings. Removal of gluten and/or casein containing products requires the active participation of all those concerned with the child's well-being. Tests have often been ruined by a well-meaning relative who ignores parental instruction or by schools or therapists who feel that the proposals are rubbish. Carers must satisfy themselves that the diet is being adhered to before any evaluation is possible. Gluten and Casein free products, together with advice on their use, are available from most good Pharmacies. Nutritionist and dieticians are also in a position to give advice. Initially the reported effects may be negative. Upset stomach, anxiety, clinginess, dizziness, aches and pains and slight ill-temper have all been reported. Experience would suggest that these are good signs and precursors of a positive response. Reichelt recommends a trial period of three months. Experience also suggests that the results are more easily demonstrated in younger children. The effects in fully grown individuals appear less impressive. It should also be noted that the withdrawal effects may also be more noticeable in small children and that these can sometimes be very marked. Where younger children are involved (less than 4 years old for example), it may be appropriate to withdraw the offending foods in stages over a period of two weeks. Given that there appear to be a number of possible causes of autism it is not unexpected that no unitary solution will be found for all cases.
Although the hypothesis may appear "off the wall" in many respects, there are a number of pieces of evidence, which seem to support them. The ideas are compatible with virtually all the accepted biological data on autism and are therefore worthy of consideration. The dietary method must still be considered as experimental and no positive results can be promised or are claimed for every person. Despite continuing scepticism about the efficacy of dietary intervention amongst some quarters of the professional community, the use of diet may well be far less harmful than other medical interventions or therapeutic regimes; although care is still necessary during its implementation. We would be pleased to receive any feedback of a positive or negative nature from anyone utilising such dietary modification in the amelioration of autism.
Additional Published References (abstract only)
Several additional references pre-1999 are present in the research literature (particularly from Prof. Knivsberg’s group in Norway). The following represent the more significant recent findings.
Whiteley (1999) A gluten free diet for autism
Knivsberg (2001) Reports on dietary intervention
Bowers (2002) Dietetic referrals
Cornish (2002) Food choice & nutrition
Knivsberg (2002) Randomised trial of diet
Arnold (2003) Nutritional intake on diet
Millward (2004) Cochrane review of dietary intervention
Ashwood (2004) Cytokine profiles on diet
Elder (2006) Double-blind trial of diet
Christison (2006) Elimination diets for autism