November 2009  << Back  

  Can't open the PDF? Click here for help.



A review of dietary interventions in autism

Pramila Srinivasan, PhD

Founder, Medical Mine, Inc. Palo Alto, CA, USA

BACKGROUND: Anecdotal reports and parent surveys have shown evidence that dietary interventions have had some success in ameliorating the symptoms of autism.

METHODS: In this paper, key findings that prompt a dietary intervention strategy are reviewed and popular intervention diets are described.

RESULTS: There is a significant body of literature pertinent to dietary interventions in autism from the perspectives of gastroenterology, immunology, and excitotoxicity. Some articles report benefits to patients on standardized rating scales.

CONCLUSIONS: This article presents a survey of the literature related to dietary interventions studied in the context of autism as well as various hypotheses on the rationale for dietary interventions. Patients or caregivers increasingly are attempting such interventions. Further studies are needed to establish the efficacy of these diets, the patients who would best benefit from diets, the mechanism of action, and the role of diets in addition to other treatments.

KEYWORDS: autism, dietary interventions, gastrointestinal



In 1943, Leo Kanner first described autism in his article “Autistic disturbances of affective contact.”1 Based on that background, in the 1950s, when the psychoanalytic theories of Sigmund Freud were in vogue, autism was attributed to “refrigerator mothers,” ie, mothers who withheld love from their infants, thereby causing a withdrawal of the infant into autism.

Today there is increasing evidence that autism is a complex, multifactorial disorder involving the brain and the body, a result of genetic vulnerabilities interacting with environmental factors.2 Autism is a developmental disorder characterized by severe abnormalities in communications, social awareness and skills, and behavior. The prognosis for most children with this diagnosis is lifelong disability

Autism was a rare disability before the 1980s. The rate of occurrence was estimated at 2 to 5 in 10,000. It is now an epidemic affecting about 1 in 150 children. The dramatic growth in the rate of occurrence cannot be explained by a change of diagnostic criteria alone or by diagnostic substitution.2 It is not purely genetic, either. Although genetics plays a role, the fact that the incidence of autism in identical twins is not 100% points to external, or environmental, factors such as epigenetics, immunologic abnormalities, infectious agents, and others.

Herbert3 states that environmental factors cause or trigger autism, implying that we have to look at the whole person and whole body, given that environmental toxins and stressors affect the whole body. This involves shifting from the model of autism as a genetically determined brain disorder to a newer and more inclusive model that considers autistic behavior one of many effects of genetic and environmental impacts on the whole person, including the brain.

In current practice, the recommended treatment for autism entails educational therapies: applied behavior analysis, speech therapy, sensory integration therapy, auditory therapy, etc. However, anecdotal reports and parent surveys have shown some evidence that diets based on food elimination and rotation, as well as supplementation and alternative treatments aimed at intestinal healing and detoxification, have had some success in ameliorating the symptoms of autism. This has prompted research into a whole-body treatment approach, with the suggestion that autism should be treated as a whole-body condition.4 Treatment approaches under recent study include supplementation, detoxification, dietary intervention, treatment of gastrointestinal (GI) symptoms, treatment of chronic inflammation in the brain and intestines, and immunologic treatments, among others.2

In particular, GI symptoms have received much attention. The aim of this article is to survey the research pertinent to the use of dietary interventions to treat autism, and to describe specific interventions that are used either alone or as a useful adjunct to other therapies. The following conditions are described in the context of autism, and we provide insights into why each of these can possibly be addressed by dietary interventions: chronic inflammatory processes, immunologic dysregulation, toxic overload, and GI pathologies.

The following section reviews those symptoms of autism that might be treatable using dietary strategies. The next section describes the special diets, the principles that support their use in autism, and the implementation strategies used.

Autism symptoms that might be treated with dietary interventions

Gastrointestinal symptoms. A number of studies have attempted to present a link between GI symptoms and autism. A few key articles are reviewed in this section. Afzal et al4 studied abdominal radiographs of 103 autistic children and 29 controls. They reported that moderate or severe constipation occurred more frequently in patients with autism (about 36%) compared with controls (10%). Analysis of dietary data showed consumption of milk to be the strongest predictor of constipation in the autism group. The article concludes that constipation is a frequent and significant problem in autism (often in association with megarectum) and continues to remain underrecognized.4 Effective treatment may potentially alleviate behavioral problems and disturbed sleep patterns in autistic children.

Finegold et al5 studied the GI microflora of 10 children with late-onset autism. All patients had chronic diarrhea and/or constipation and more Clostridia species than did controls. The authors also postulated that autoantibodies or another bacterial antibody reaction could lead to autism.

Ashwood et al6 surveyed mucosal immunity in autism in an effort to establish a link between GI symptoms and autism. In this review, GI symptoms were seen in 18% to 40% of patients with autism and included abdominal pain, bloating, diarrhea, and constipation, among others. Reported abnormalities included low disaccharidase enzyme activity, defective sulfation, bacterial overgrowth of clostridial species, increased intestinal permeability, and positive effects on cognition following dietary interventions. Flow cytometric analyses have shown abnormalities of lymphocyte populations at different anatomic sites: stomach, duodenum, ileum, colon, and others. Although lymphoid nodular hyperplasia (LNH) is not uncommon in patients with allergies, the increased severity and frequency of LNH seen in patients with autism, along with increased intestinal permeability, may perturb the intestinal barrier function in these patients. It is not entirely clear how mucosal changes affect behavior, but certain theories have been suggested. It may be that antigens in the diet can more readily cross the mucosa, where they cause local inflammatory reactions, generating proinflammatory cytokine signals that interact with afferent neurons. Also, failure to detoxify neuroactive antigens may lead to cognitive impairments. Similarly, passage of opioid peptides from the diet into the body may cause behavioral abnormalities. Another theory postulates that reduced absorption of vitamin B12, a necessary factor in the formation of myelin, may impair nerve function.6 Conclusive evidence demonstrating how the disrupted mucosa could affect the CNS is still needed. In a study by Valicenti-McDermott and colleagues, 7 a history of GI symptoms was elicited in 70% of children with autism compared with 28% of controls.

The relationship between food allergies and infantile autism was explored by Lucarelli et al8 in a study of the effect of a diet free of cow’s milk (and other foods giving a positive skin test) in 36 patients with autism.8 Skin tests were carried out on all patients using prick tests. For 8 weeks, the autistic patients followed a diet free of cow’s milk and other foods that caused a positive skin test. A test using an autistic behavior evaluation scale (BES) was performed prior to and after the 8-week diet. In those cases in which an improvement was observed, a double-blind, placebo-controlled challenge with the food allergens was performed. The behavioral pattern was assessed after 2 weeks. Total IgE and serum levels of IgG, IgA, and IgM were determined, which were specific for cow’s milk and egg proteins. A significant improvement in behavioral disturbances was achieved in patients in 5 of the 7 groups on the BES scale, which was applied before and after the diet. The skin prick test was positive in 36% of the autistic patients and 5% of the controls. High levels of IgA-specific antibodies for casein, lactalbumin, and IgG and IgM for casein were noted. The authors concluded by hypothesizing that there is a relationship between food allergies and infantile autism.8

More details relating GI symptoms to possible abnormalities in immune response are provided in a study by Jyonuchi et al.9 In this work, the authors described dysregulated innate immune responses in young children with autism spectrum disorders (ASD), their relationship to GI symptoms, and dietary interventions. Innate immunity mounts the initial nonspecific immune response in an antigen-independent manner, but proper innate responses are needed for effective adaptive immunity and are closely associated with CNS functions. A previous study10 found an association between cellular immune reactivity to common dietary proteins (DPs) and excessive proinflammatory cytokine production with endotoxin, a major stimulant of innate immunity in the gut mucosa in a subset of autistic patients. It remained to address whether such abnormal responses are intrinsic to these autistic patients or whether the results of chronic GI inflammation are secondary to immune reactivity to DPs. These findings indicated intrinsic defects of innate immune responses in patients with GI symptoms (GI-positive), suggesting a positive link between GI and behavioral symptoms that is mediated by innate immune abnormalities. This study revealed elevated TNF-alpha production with lipopolysaccharides both in GI-positive and GI-negative (absence of GI symptoms) patients with autism, and in both groups, one with diet restriction and one without. However, the levels were higher in the unrestricted diet group. This suggests an intrinsic dysregulated innate response in GI-positive autistic patients, which may pre-dispose these individuals to adverse reactions to dietary proteins and aggravation of behavioral symptoms. The study authors concluded that GI-positive patients with autism have complex defects of innate immunity, but a better understanding of these immune defects to CNS functions is needed.

Knivsberg et al11 describe a single-blind study that evaluated the effect of a gluten- and casein-free diet on a small group of 20 children. The children all had an autism diagnosis and urinary peptide abnormalities. The authors’ hypothesis was that the products of incomplete breakdown of proteins could result in caseomorphines and gluteomorphines. These compounds could have opioid-like effects if absorbed through a permeable intestine. Evaluation tests were performed before and after a period of 1 year in the areas of communication, language, and motor and social skills. The authors noted that development was significantly better in the group of children following the gluten- and casein-free diet. Improvements were noted in social connection, willingness to learn, ability to make transitions, and other areas. The children following the diet had fewer autistic traits after this 1-year period.11

O’Banion et al11 studied the response of an 8-year-old patient with autism to normal diet, then fasting, followed by careful presentation of test foods to observe behavioral responses; the observation period was approximately 1 month. The responses included incidents of hyperactivity, uncontrolled laughter, screaming, biting, scratching, and object throwing. The discussion illustrated how certain foods increased negative behaviors and how the subject functioned well on a diet of nonreactant organic foods, with no severe episodes of disruptive behavior. Goodwin et al13 studied various interrelationships, including malabsorption and cerebral dysfunction, in 15 children with autism and found that gliadin had profound effects on blood cortisol levels. The authors concluded that this may indicate that autism is a disorder in which an underlying cerebral defect may be aggravated by a normal chemical response to stress and may be influenced by dietary factors. Such responses suggest a correlation of autism with malabsorption and sensitivity to food.

Based on endoscopies of hundreds of pediatric patients, Krigsman14 described in detail the various lesions found in patients with autism. The author noted that many patients responded well to some combination of restrictive diets, anti-inflammatories, probiotics, enzymes, antibiotics, and antifungals. The author described LNH, or enlarged clusters of immune cells, at the back of patients’ throats, which occurred in response to an immunologic trigger (eg, food allergens, viruses, bacteria). The author suggested that several posturings, tantrums, or other behaviors could result from GI pain. He also noted findings of inflammatory polyps in these patients, which were not precancerous but, rather, resulted from stomach inflammation. A pill-camera procedure on the small intestinal area often shows polyps, ulcers, erosions, and LNH in patients with autism. Similarly, the author reported observing inflammatory polyps, ulcerative colitis, Crohn’s disease, and LNH in the colon.

Torrente et al15 compared duodenal biopsies of 25 children with regressive autism with children with celiac disease, cerebral palsy, and controls. The authors concluded that there may be a novel form of enteropathy in children with autism, with increased lymphocyte density and complement deposition, suggesting an association between autoimmune lesions and autism.

The basic gluten- and casein-free diet (which will be further described in the next section) is often the first dietary intervention attempted by parents of children with autism. As stated previously, this diet requires the elimination of gluten (commonly found in wheat, oats, barley, rye, and processed foods) and casein (found in milk products). This intervention addresses the possibility of passage of opioid peptides from the diet into the CNS. Some of the dietary interventions described later, such as the Specific Carbohydrate Diet, are designed to heal intestinal permeability and combat overgrowth of pathogens in the gut mucosa. The use of anti-inflammatory foods and antioxidant-rich foods is also a core principle in such dietary interventions.16,17 Elemental diets and other cleansing strategies may also be performed under the supervision of a physician.

Immune dysfunction and inflammation in autism and the role of diets in lowering immune activation. Ashwood and de Water6 reviewed immunologic dysfunctions in autism, including effects on neurotransmitters, autoimmune processes, and aspects of mucosal immunity. Neurologic findings included a reduced number of neurons, increased brain size, and decreased numbers of Purkinje cells in autism. It has been proposed that increased and abnormal levels of neuroactive compounds may lead to autism.6,18 These findings suggest a role for CNS inflammation in autism.

Autoimmune processes have also been noted in autism, eg, the presence of antibodies directed against components of the CNS in the sera of autistic children. Various antibrain antibodies have been found, including those against serotonin receptor, myelin basic protein, neuron axon filament protein, nerve growth factor, and others. The exact signature mechanisms are conflicting, but taken together with findings of autoimmunity in families, they suggest that antibodies that target the CNS may be a pathologic or exacerbating factor in children with autism.6

Many infectious agents, including rubella, measles, human herpesvirus 6 (HHV-6), influenza, and cytomegalovirus, have been associated with the etiology of autism.19,20 This association may highlight an underlying inability to fully eradicate viral insults. Persistent viral infection has been described in the mucosal tissue of autistic children and may be reflective of a local inflammatory response to the presence of a persisting antigen. Patients with antimeasles IgG antibodies were also positive for anti–myelin basic protein (MBP) and anti–neuron-axon filament protein (NAFP) antibodies.21 The underlying mechanism between viral insults and autism is unclear,6 but it may result from cellular damage induced as a result of immune responses aimed at eradicating an invading virus. In addition, Singh et al22 have reported on autoimmunity to the CNS, especially to MBP.

Neuroimmune studies in autism suggest innate immune responses and inflammation. In 2005, Pardo et al23 published a study on immunity, neuroglia, and neuroinflammation in autism. This is an important neuroimmunopathologic study, which suggests that innate, rather than adaptive, neuroimmune responses are among the immunopathogenic mechanisms associated with autism; however, this does not exclude other cellular or humoral responses at early stages of the disease. The Pardo et al23 article demonstrated the presence of neuroglial and innate neuroimmune system activation in brain tissue and cerebrospinal fluid (CSF).

Neuroglial cells such as astrocytes and microglia, along with macrophages, play important roles in neuronal function and contribute to the regulation of immune responses in the CNS. Evidence of neuroglial activation and a role for neuroimmune responses mediated by innate immunity was reported in this article. Based on neuropathologic analyses of postmortem brain tissue from 11 autistic patients, the authors demonstrated the presence of an active and ongoing neuroinflammatory process in the cerebral cortex and white matter and showed marked activation of astroglia and microglia. Marked microglial activation was found in the cerebellum and cortical and white matter regions. Such CNS activation may play a role in the inflammatory responses of the brain.23

Although the effect of dietary approaches on brain inflammation was not directly studied, this research, reported by Guaner and Malagelada24 prompts us to look into all therapies known to suppress inflammation. Probiotic foods are living microorganisms that, upon ingestion in specific numbers, exert health benefits beyond those of inherent nutrition.24 Resident bacterial flora may be a factor in driving inflammatory processes in bowel diseases, and gut flora have a function in the development and homeostasis of the immune system as well as protection against pathogens. Probiotics do not necessarily colonize the human intestine, and well-designed clinical trials are needed to demonstrate probiotic activity.24 Fermenting also produces easier-to-digest foods rich in lactic acid, an aid to a weak digestive system. Fermented foods are also raw and rich in enzymes. For this reason, many of the immune-supportive diets described below recommend traditional foods such as cultured vegetables and kefirs.

In their 2006 article, Olivares et al25 described the effect of deprivation of dietary fermented foods on the immune system in 30 healthy adult volunteers and found significant decreases in fecal lactobacillus and total aerobes count as well as the concentration of short-chain fatty acids. Moreover, they noted a decrease in phagocytic activity in leukocytes after 2 weeks of a restricted diet. The authors concluded that dietary deprivation of fermented foods could induce a decrease in innate immune response that might affect the capacity to respond against infection. The ingestion of a probiotic product containing the strains Lactobacillus gasseri CECT5714 and Lactobacillus coryniformis CECT5711, or a standard yogurt containing a conventional starter, Lactobacillus delbrueckii subsp. Bulgaricus, counteracted the decline in the immune response, although the probiotic product was more effective than the standard yogurt. The authors concluded that the ability of the Lactobacillus strains to counteract the effects of deprivation of fermented foods demonstrated the role that probiotic-enriched foods might play in the diet.

There is also evidence of the benefit of elimination and rotation diets based on IgE and IgG food and mold panels. Drisko et al26 noted that in irritable bowel disease (IBD), the gut-associated immune system may be upregulated, resulting in immune complex production, low-grade inflammation, and translocation of inflammatory mediators and macromolecules outside the GI lumen. Since food intolerance may be one of the reasons for this upregulation, the goal was to study the effect of food intolerances. This 1-year study enrolled 20 patients, who underwent baseline testing (IgE and IgG food panels, stools analysis, etc.). The patients followed food-elimination diets and used probiotics. Significant improvement was reported in GI condition (based on stool tests and a repeat of baseline tests) after a year of elimination and rotation diets. There was also an increase in beneficial flora. A significant number of participants continued the diet after the completion of the trial.

Similarly, Atkinson et al27 showed a significant improvement in IBD symptoms after a 3-month trial of 150 patients who followed a food elimination diet based on IgG tests. The authors pointed out that the dietary strategy is worthy of future research.

Abnormal intestinal flora, pathogens, and the role of cultured foods. Finegold et al5 study GI microflora in 10 children with late-onset autism who had chronic diarrhea and/or constipation. Polymerase chain reaction (PCR) tests revealed 25 species of Clostridia in stool specimens of autistic individuals and 15 species in the control group. The patients responded to oral vancomycin, although relapses were common. The 2 most reasonable hypotheses for this colonization were impaired gastrointestinal motility and IgA deficiency.

The role of Candida in the symptoms of autism has been strongly suspected among patients pursuing biomedical interventions for autism. Candida yeast produces toxins that could cause some of the symptoms directly or by overloading the detoxification pathways. Shaw and colleagues’ article on antifungal treatments in autism described a study involving 23 children with autism and a control group.28 The study found that antifungals lowered specific key markers in urinary organic acid tests that might be byproducts of Candida over-growth. Diets targeting yeast overgrowth range from low-carbohydrate and probiotic-rich diets, such as the Specific Carbohydrate Diet and the Body Ecology Diet, to yeast-free diets.

Excitotoxicity and the role of diets. Blaylock29 described the process of nerve cell stimulation by glutamates and aspartates, and showed that monosodium glutamate (MSG) is a potent cytotoxin that can kill brain cells. Glutamates and aspartates can cause neurons to fire their impulses very rapidly. In large enough doses, they can cause cells to degenerate and die. Even small doses, according to the author, can cause nerve damage. Neuroscientists call this class of chemicals “excitotoxins.” Although Blaylock’s book addresses adult degenerative diseases, similar references to glutamate toxicity are made in the context of autism. Shinohe et al30 found increased serum glutamate levels among adult autistic patients.

Many caregivers provide patients with diets free of MSG and artificial colors or flavorings in order to address their possible relationship with autism. The Feingold Diet also recommends against the use of MSG in the context of hyperactivity. McCann and colleagues31 reported on a randomized, double-blind, placebo-controlled study on the effect of food colorings and additives on childhood behavior. Children were between age 3 (n=153) and 9 (n=144) and were drawn from the general population. Three measures of hyperactivity were used to calculate a global hyperactivity aggregate: abbreviated ADHD rating scale IV (teacher’s edition), abbreviated Weiss-Werry Peters hyperactivity scale, and the classroom observation code. For children age 8 to 9, the Conners Continuous Performance Test II using visual stimuli was performed. The authors concluded that sodium benzoate preservatives and/or food coloring or artificial additives in the challenge drink resulted in a significant increase in hyperactivity.

Mitochondrial dysfunction, oxidative stress, and antioxidant diets. Classic mitochondrial diseases occur in a subset of children with autism and are usually caused by genetic anomalies. However, in many cases of autism, there is evidence of mitochondrial dysfunction without the classic features, which presents less severe symptoms. However, an article by Rossignol and Bradstreet states that this dysfunction might contribute to symptoms of autism, such as cognitive impairment, language deficits, increased oxidative stress, and others.32 During mitochondrial respiration, the inability to neutralize reactive oxygen species and free radicals leads to oxidative stress. Oxidative stress is known to contribute to aging and neurodegenerative disease in humans. Impairments in mitochondrial function could lead to further oxidative stress and lower glutathione levels, and a vicious cycle may ensue.

MacInnis investigated the role of oxidative stress in autism.33 Impaired energy production and oxidative stress are also intimately related to excitotoxicity. Greater oxidative stress increases release of glutamates. In addition, in patients with mitochondrial disease, increased ammonia has been observed. An important dietary strategy against excitotoxity would be the avoidance of dietary excitotoxins such as MSG and aspartame in food and drink.34

An antioxidant-rich dietary intervention is a possible strategy to lower oxidative stress and perhaps dampen the vicious cycle involving mitochondrial dysfunction and oxidative stress. A dietary strategy for lowering ammonia in such patients is to consume lower amounts of protein, such as meat. Many industrial toxins, including pesticides, can inhibit mitochondrial function. Again, a diet high in antioxidants and organic raw and fresh fruits and vegetables seems appropriate.

Another area that has been studied that lends itself to dietary intervention is the presence of abnormalities in sulphur metabolism in autism, as evidenced by reduced levels of plasma sulphates. Waring and colleagues35 explain the role of sulphotransferase enzymes in the sulphation of phenols and amines. The amines in bananas, chocolate, cheese, etc., are affected by the reduced enzyme activity. Compounds such as flavonoids in citrus fruits inhibit the enzyme. The combination of low enzyme activity and low sulphate availability greatly reduces the capacity to detoxify amines and phenols. The dietary recommendation for this condition is—apart from gluten and casein removal—the removal of chocolate, bananas, citrus fruits, vanillin, and food colorants. Another approach is to lower the intake of sulfur-containing foods, such as cabbage, broccoli, egg yolks, etc.

A review of dietary approaches to address the underlying conditions in autism

In most cases, the dietary approaches described below were developed for conditions related to autism, such as inflammatory bowel conditions, multiple food sensitivities, Candida and viral infections, etc. As such, they were adapted to autism to address similar conditions in patients. Following is a brief description of the diets, with a discussion of the related conditions seen in autism and the dietary aspects of interventions.

Food elimination or rotation diet. Some patients with autism have been shown to produce excess levels of IgE and IgG antibodies in response to allergens in food, medicine, and environmental inhalants.2,8,15,36 Trajkovski et al36 identified specific IgA, IgE, and IgG antibodies to food antigens in 35 participants with autism and 21 of their siblings. Statistically significantly higher plasma concentration of IgA antibodies against alpha-lactalbumin, beta-lactoglobulin, casein, and gliadin were found in children with an ASD. Plasma concentrations of IgG antibodies against alpha-lactalbumin, beta-lactoglobulin, and casein in participants with autism were significantly higher. IgE-specific antibodies (alpha-lact-albumin, beta-lactoglobulin, casein, and gluten) as well as total IgE were also statistically significantly higher in patients with autism.36

IgE antibodies result in a histamine release, which leads to symptoms that are commonly recognized as allergic reactions. Children with autism tend to have IgE-mediated sensitivities2 as part of their immune dysregulation syndrome. The dietary aspect of the intervention for this condition is the testing and removal of food items that the person is sensitive to.

Although IgG-mediated responses are not associated with immediate visible symptoms, they might play a role in eczema, headaches, and other conditions—a possible indicator of immune system reactivity.2 Although there is ongoing debate about the relative importance of this response, one of the dietary interventions being investigated by parents and caregivers is the elimination and/or rotation of foods that are identified as highly sensitive on IgG food-sensitivity tests. The typical dietary strategy is to eliminate the most reactive foods, to rotate less reactive foods on a daily basis, and to primarily consume nonreactive foods. This strategy has resulted in reduced food sensitivities overall, and there is an indication that clinical symptoms of autism improve as a result.2

Gluten and casein elimination diet

A diet that eliminates gluten and casein is often the starting point in the journey of dietary interventions. This simple diet is a foundation for some of the more advanced diets. As stated earlier, the gluten and casein elimination diet calls for the complete removal of both gluten, which is found in wheat, rye, barley, and oats, as well as casein, the protein in milk and all milk products. The elimination of milk products is sometimes complicated by the fact that some patients experience a deep craving for cow’s milk. Again, a variety of options are available to substitute cow’s milk with rice/hemp milk, nut milks, and so on.

Many parents consider the gluten and casein elimination diet a very effective intervention and report improvements at various levels: eye contact, connection, and others. Several hypotheses have been suggested to explain these improvements. Elder et al37 commented on notable reports from parents and teachers of children being “cured” of autism by the implementation of gluten- and casein-free diets, having acquired language and marked social connectedness. Their article describes the design of a double-blind, controlled trial of the gluten-free and casein-free diet in 15 children, age 2 to 16 years. The foundation for this diet is drawn from schizophrenia, where it was asserted that patients, possibly due to a genetic defect, suffered an overload of gluten and milk proteins (casein). Normally, the proteins would be metabolized into peptides and then into amino acids, which would be absorbed by capillaries in the intestines. High peptide levels may be caused by intestinal permeability. The article also reported on a prior study on 149 children diagnosed with autism who underwent this dietary intervention, with 81% of the children reported as showing improvement in 3 months. However, the authors commented that most assessments were made by parents who were aware that the children were on the diet. Knivsberg et al11 reported on a single-blind study of a gluten and casein elimination diet, using 10 children each in the patient and control groups. Observations and tests were performed before and after 1 year. Tests were conducted using non-verbal techniques, and other linguistic tests were also performed. Statistically significant improvements were found in the diet group with respect to aloofness, routines and rituals, and responses to learning.

Vojdani, Pangborn et al38 measured the antibodies IgG, IgM, and IgA against gliadin, casein, brain myelin basic protein, egg, corn, and soy in 50 children diagnosed with autism. Blood sample analysis indicated that a significant number of children developed antibodies against casein and gliadin. In addition, casein and gliadin were shown to bind to the lymphocyte and tissue enzyme CD26 and were thought to trigger inflammation and an immune response.37

The discussion above indicates that more than one source supports the hypothesis justifying the trial of a gluten-free and casein-free diet for autism. Elder and colleagues37 describe a study to evaluate the benefits of this diet using the Childhood Autism Rating Scale (CARS) to measure the benefit of the diet. Urinary peptide levels were also measured, and the parents’ role in the diet was also studied. The participants received a supply of compliant food and instructions on what to use if they did not have access to their supply of compliant food. In the study group analysis, no statistically significant results were found in CARS scores or urinary peptide levels. However, anecdotal reports were interesting: Decreased hyperactivity and tantrums and increased language were reported in 7 children. A similar positive report was provided by a teacher and a respite worker. The authors commented that this could be due to the small sample size and the heterogeneity of the severity, cognitive levels, and symptoms in the population chosen and called for more extensive studies with more homogeneous subgroups to learn about the possible benefits of the diet.

Although the gluten and casein elimination diet may be useful, several patients may need further refinements to this diet and may adopt refinements from other dietary approaches.

Specific Carbohydrate Diet. The Specific Carbohydrate Diet is a natural diet that was pioneered by Elaine Gotschall39 as a treatment for ulceritive colitis. She also popularized the diet among the autism community as a means of treating GI symptoms and possibly behavioral issues in autism. The carbohydrates in this diet are predominantly single sugars (monosaccharides) such as those found in fruits, honey, properly made yogurt, and certain vegetables. In the book Breaking the Vicious Cycle, Gotschall describes the cycle that is established as a result of improper digestion of complex carbohydrates. Carbohydrates that escape digestion and, therefore, absorption remain in the intestinal tract and are used by the microbial world of the intestine. Overgrowth of yeast and bacteria can result in dysbiosis, bacterial and fungal by-products, and production of mucus; it can also injure the intestine, which may in turn contribute to impaired digestion of disaccharides—thus, a vicious cycle ensues. The Specific Carbohydrate Diet is based on the premise of starving out these organisms and reestablishing intestinal health.

The foods allowed on this diet are proteins such as meats, eggs, natural cheese, homemade yogurt, nonstarchy vegetables such as cabbage, cauliflower, onions, spinach, and peppers, and a variety of nut and nut flours such as almonds, brazil nuts, and walnuts, and soaked lentils and beans. No grains or products with trace amounts of grains may be consumed.

Body Ecology Diet. The Body Ecology Diet, pioneered by Donna Gates,16 focuses on rebuilding immunity and intestinal flora. This diet uses traditional healing principles such as the use of fermented foods (eg, young coconut kefir and cultured vegetables). Another important aspect of this diet is the establishment of a slightly alkaline pH to enable beneficial flora to flourish, using principles of food combining and the use of alkaline-forming foods. The benefit of this approach is supported by studies on intestinal flora in autism.5 By advocating the use of probiotic-rich fermented foods and the slightly alkaline food balance, this diet addresses Candida overgrowth by limiting carbohydrates and promoting beneficial gut flora. Vargas et al40 limited simple carbohydrates in a mouse model and observed reduced Candida colonization.

Similar to macrobiotic diets, the Body Ecology Diet recommends the use of seaweeds and grains such as buckwheat. Other grains used in this diet are quinoa, millet, and amaranth. The grains are soaked and fermented, sometimes sprouted, to allow for easy digestion. Every meal is balanced on a 80-20 rule, balancing starchy vegetables, nonstarchy vegetables, and proteins. Nonstarchy vegetables make up the bulk of the diet, with some room for starchy vegetables, proteins, and grains. Dairy products are not encouraged initially, but the use of raw butter is investigated at a later stage. Although the Body Ecology Diet is difficult to implement, it is a balanced diet that returns our attention to the power of traditional foods, and the use of fermented foods to establish healthy gut flora.

Feingold and Failsafe Diets. The Feingold Diet is described in the book Why Your Child Is Hyperactive.41 It is based on the benefit of a food-restriction diet for attention-deficit and hyperactivity disorders. The author, Ben Feingold, provides several anecdotes of behavioral improvements after the elimination of food colors and flavors. In Feingold’s observation, elimination of food additives caused some of his hyperactive patients’ symptoms to decline so dramatically that, in some cases, they could be weaned off of medications. Apart from synthetic substances, he also recommended avoiding natural salicylates, which are found in a number of fruits and in tomatoes and cucumbers. As such, Group I foods contain natural salicylates, and Group II foods contain synthetic additives. Some of the items in Group I are almonds, apples, apricots, berries, cherries, grapes, nectarines, oranges, peaches, plums, and prunes. Group II items are most factory-made cereals, instant breakfasts, baked goods, luncheon meats, frozen fish, candies, soft drinks, and desserts or other foods that contain synthetic additives for color or flavor. Feingold advocates avoiding these foods and maintaining a detailed dairy to record the observed behavior that follows use of these diets.

Hyperactivity is often a condition that occurs in ASD,42 and parents may adopt this diet as an intervention to possibly help with attention and hyperactivity issues. Salicylates are a subgroup of phenols, and some parents notice that patients with autism have occasional problems with breakdown of phenols. A deficiency of phenol sulfur transferase (PST) in autism has been noted; the PST enzyme is involved in the breakdown of phenols and amines.35 Anecdotally reported symptoms of salicylate sensitivities in children include dark circles under the eyes, hyperactivity, difficulty falling asleep at night, inappropriate laughing/giggling, head banging, and others. Some of the high-phenolic foods that these patients should avoid include tomatoes, apples, peanuts, bananas, oranges, some red- and orange-colored fruits, and foods on the Feingold list. Other exposures that might be avoided for this reason include perfumes and fragrances and MSG. Some enzyme products to help the processing of phenols are available.

Another variation on the principle of avoiding food additives is the Failsafe Diet, described by Sue Dengate,43 as an intervention for ASD. This diet advocates the reduction of food additives, salicylates, amines, and flavor enhancers such as MSG.

Raw food and antioxidant diets. Raw food diets are based on uncooked and unprocessed foods such as sprouts, fresh fruits and vegetables, dried fruits, seeds, nuts, grains, and seaweed. Heating food above 116°F is thought to destroy the enzymes in raw foods, so this diet is designed to preserve enzymes in the foods. This diet recommends soaking of seeds and nuts and using a dehydrator to prepare foods, with the goal of providing nutrient-dense foods that retain their natural enzymes and antioxidants.

Antioxidant diets are of interest to those who are concerned with oxidative stress in autism and other conditions. The recommended food groups are fresh vegetables, fresh fruits, cooked legumes, starchy vegetables, and whole grains. Moderate servings of animal products such are lean meat are allowed. The super foods in this diet are broccoli, brussels sprouts, berries such as blueberries and cranberries, Goji berries, and sea vegetables.17

As discussed above, inflammation is studied in the context of autism, and it is also the cause in other inflammatory diseases caused by an overactive immune system, such as arthritis, allergies, asthma, lupus, and others. Dietary strategies to treat these conditions could prove beneficial in autism as well. A major culprit is the production of excess arachidonic acid, an omega-6 fatty acid. Unlike roughage-fed animals, grain-fed animals are thought to produce excess arachidonic acid,44 so limiting or eliminating farm-raised fish, chicken, and animals from the diet is recommended. Consumption of omega-3 and omega-6 in a 1:1 ratio is also recommended. The nightshade family vegetables, such as white potatoes, tomatoes, and eggplant, are also limited because they produce solanine, a chemical that causes inflammation.

Ketogenic diet. The ketogenic diet, developed about 80 years ago for patients with seizures, relies on increased fats and a smaller proportion of carbohydrates. The ketogenic diet requires the supervision in a clinical setting.

Evangeliou and colleagues45 examined the efficacy of a ketogenic diet in autism in a pilot study of 30 children. The John-Radcliffe version of the diet was chosen for this study. The energy intake was distributed as 30% medium-chain triglyceride oil, 30% fresh cream, 11% saturated fat, 19% carbohydrate, and 10% protein. Patients were also given vitamin and mineral supplements, dosed according to age. The children were evaluated using CARS scores before and after the dietary intervention. Significant improvement was seen in 2 patients (12 units on the CARS scale), average improvement in 8 patients (8 to 12 units), and minor improvement in 8 patients (2 to 8 units). The authors concluded that the ketogenic diet should be investigated further as an alternative or additional treatment for autism.

General strategies in dietary management. In their study of 30 participants with autism or pervasive developmental disorder not otherwise specified (PDD NOS), Ahearn et al46 found that more than half exhibited overall low levels of food acceptance.46 The authors studied the factors that might have influenced the food acceptance: texture, food group, etc. They suggested that the food selectivity might be caused by a common behavior of restricted interest in patients with autism, and that food acceptance may be a cause of stress in families. In another study, the lead author47 reported successful strategies for increasing levels of acceptance for nonpreferred foods (in this case, vegetables) by using condiments and successful presentation of preferred and nonpreferred items. Further studies are needed to investigate the causes of feeding problems in children with autism. Possibilities include anxiety about unknown/unfamiliar foods; craving for certain foods; dislike of certain foods due to GI upset; loss of appetite; sensations such as acidity, bitterness, etc., in the mouth; deficiencies that might affect the sense of taste and sensory perception; and others. In one study, Schreck et al48 reported that food acceptance was correlated to oral motor issues, use of certain utensils, etc. A related article described a study of 298 children with an ASD and a control group of 138 children recruited from local schools to complete a questionnaire.34 The results were statistically tallied to determine food preferences and the nature of preferences, such as types of utensils, textures, types of presentations, etc., and it was found that the patients with an ASD ate a significantly smaller set of foods in each food group. The authors concluded that such eating differences should be addressed to prevent nutritional deficiencies in children with ASD.

The preceding discussion on dietary approaches suggests that there are contradictions in dietary philosophies (eg, a good food in one diet is avoided in another diet). In most cases, significant trial and error and careful record keeping are needed to develop a dietary strategy that might work for an individual patient. In some cases, the dietary strategy may involve reduction or rotation, rather than complete elimination, of a food. In that case, the rotation is implemented, and overall symptoms are noted. If general improvement in symptoms is observed, the rotation might be beneficial. Rotation is often used in conjunction with IgG tests and other assays, with the help of physicians, nutritionists, or other professionals.


Parents with children diagnosed with autism now have access to a variety of options to try alleviate their children’s symptoms. There is some evidence emerging that GI distress and immunologic dysfunction could contribute to several of the behavioral and perhaps cognitive symptoms in autism. Parents have access to medical research, Internet-based forums, parent blogs, and Web sites, and, in short, several sources of anecdotal data about improvement in symptoms of autism as a result of dietary interventions. It is inevitable that parents will attempt these interventions with or without the help of a nutritionist or other qualified practitioner, at least as a trial intervention. There is a need for more standardized protocols for recommendations of dietary interventions, given patients’ requirements and concurrent treatments. Each patient’s nutritional needs, responses to various interventions, laboratory tests, concurrent treatments, practical concerns in implementation, and family diet are all possible factors in an effective dietary strategy.

DISCLOSURES: Dr. Srinivasan is the founder of Medical Mine, Inc., a software service to manage autism treatments.

ACKNOWLEDGEMENTS: The author acknowledges several detailed and insightful comments by the reviewers.


  1. Kanner L. Autistic disturbances of affective contact. Acta Paedopsychiatr. 1968;35:100–136.
  2. Jepson B, Johnson J. Changing the course of autism. Boulder, CO: Sentient Publications; 2007.
  3. Herbert MR. Autism: a brain disorder or a disorder that affects the brain? Clinical Neuropsychiatry. 2005;2:354–379.
  4. Afzal N, Murch S, Thirrupathy K, et al. Constipation with acquired megarectum in children with autism. Pediatrics. 2003;112:939–942.
  5. Finegold SM, Molitoris D, Song Y, et al. Gastrointestinal microflora studies in late-onset autism. Clin Infect Dis. 2002;35:S6–S16.
  6. Ashwood P, Van de Water J. A review of autism and the immune response. Clin Dev Immunol. 2004;11:165–174.
  7. Valicenti-McDermott M, McVicar K, Rapin I, et al. Frequency of gastrointestinal symptoms in children with autism spectrum disorders and association with family history of autoimmune disease. J Dev Behav Pediatr. 2006;27:S128–S136.
  8. Lucarelli S, Frediani T, Zingoni AM, et al. Food allergy and infantile autism. Panminerva Med. 1995;37:137–141.
  9. Jyonouchi H, Geng L, Ruby A, et al. Dysregulated innate immune responses in young children with autism spectrum disorders: their relationship to gastrointestinal symptoms and dietary interventions. Neuropsychobiology. 2005;51:77–85.
  10. Jyonouchi H, Sun S, Itokazu N. Innate immunity associated with inflammatory responses and cytokine production against common dietary proteins in patients with autism spectrum disorder. Neuropsychobiology. 2002;46:76–84.
  11. Knivsberg AM, Reichelt KL, Hoien T, et al. A randomized, controlled study of dietary intervention in autistic syndromes. Nutr Neurosci. 2002;5:251–261.
  12. O’Banion D, Armstrong B, Cummings RA, et al. Disruptive behavior: a dietary approach. J Autism Child Schizophr. 1978;8:325–337.
  13. Goodwin MS, Cowen MA, Goodwin TC. Malabsorption and cerebral dysfunction: a multivariate and comparative study of autistic children. J Autism Child Schizophr. 1971;1:48–62.
  14. Krigsman A. Gastrointestinal pathology in autism: description and treatments. Medical Veritas. 2007;4:1522–1530.
  15. Torrente F, Ashwood P, Day R, et al. Small intestinal enteropathy with epithelial IgG and complement deposition in children with regressive autism. Mol Psychiatry. 2002;7:375–382.
  16. Gates D. The body ecology diet. Bogart, GA: Body Ecology; 2006.
  17. Jeep R, Couey R, Pitman Ellington S. The super-antioxidant diet and nutrition guide. Charlottesville, VA: Hampton Roads Publishing Company; 2008.
  18. Nelson KB, Grether JK, Croen LA, et al. Neuropeptides and neurotrophins in neonatal blood of children with autism or mental retardation. Ann Neurol. 2001;49:597–606.
  19. Chess S, Fernandez P, Korn S. Behavioral consequences of congenital rubella. J Pediatr. 1979;93:699–703.
  20. Binstock T. Intra-monocyte pathogens delineate autism subgroups. Med Hypotheses. 2001;56:523–531.
  21. Singh VK, Lin SX, Yang VC. Serological association of measels virus and human herpes virus-6 with brain autoantibodies in autism. Clinical Immunology and Immunopatholgy. 1998;89:105–108
  22. Singh VK, Lin SX, Newell E, et al. Abnormal measles-mumps-rubella antibodies and CNS autoimmunity in children with autism. J Biomed Sci. 2002;9:359–364.
  23. Pardo CA, Vargas DL, Zimmerman AW. Immunity, neuroglia and neuroinflammation in autism. Int Rev Psychiatry. 2005;17:485–495.
  24. Guarner F, Malagelada JR. Gut flora in health and disease. Lancet. 2003;361:512–519.
  25. Olivares M, Paz Díaz Ropero M, Gómez N, et al. Dietary deprivation of fermented foods causes a fall in innate immune response. Lactic acid bacteria can counteract the immunological effect of this deprivation. J Dairy Res. 2006;73:492–498.
  26. Drisko J, Bischoff B, Hall M, et al. Treating irritable bowel syndrome with a food elimination diet followed by food challenge and probiotics. J Am Coll Nutr. 2006;25:514–522.
  27. Atkinson W, Sheldon TA, Shaath N, et al. Food elimination based on IgG antibodies in irritable bowel syndrome: a randomised controlled trial. Gut. 2004;53:1459–1464.
  28. Shaw W, Kassen E, Chaves E. Assessment of antifungal drug therapy in autism by measurement of suspected microbial metabolites in urine with gas chromatography—mass spectrometry. The Clinical Practice of Alternative Medicine Magazine. 2000;1:15–26.
  29. Blaylock RL. Excitotoxins: the taste that kills. Albuquerque, NM: Health Press; 1996.
  30. Shinohe A, Hashimoto K, Nakamura K, et al. Increased serum levels of glutamate in adult patients with autism. Prog Neuropsychopharmacol Biol Psychiatry. 2006;30:1472–1477.
  31. McCann D, Barrett A, Cooper A, et al. Food additives and hyperactive behavior in 3-year-old and 8/9-year-old children in the community: a randomised, double-blinded, placebo-controlled trial. Lancet. 2007;370:1560–1567.
  32. Rossignol DA, Bradstreet JJ. Evidence of mitochondrial dysfunction in autism and implications for treatment. American Journal of Biochemistry and Biotechnology. 2008;4:208–217.
  33. Chauhan A, Chauhan V. Oxidative stress in autism. Pathophysiology. 2006;13:171–181.
  34. Schreck KA, Williams K, Smith AF. A comparison of eating behaviors between children with and without autism. J Autism Dev Disord. 2004;34:433–438.
  35. Waring RH, Klovrza LV. Sulphur metabolism in autism. Journal of Nutritional and Environmental Medicine. 2000;10:25–32.
  36. Trajkovski V, Petlichkovski A, Efinska-Mladenovska O, et al. Higher plasma concentration of food-specific antibodies in persons with autistic disorder in comparison to their siblings. Focus Autism and Other Developmental Disabilities. 2008;23:176–185.
  37. Elder JH, Shankar M, Shuster J, et al. The gluten-free, casein-free diet in autism results of a preliminary double blind clinical trial. J Autism Dev Disord. 2006;36:413–420.
  38. Vojdani A, Pangborn JB, Vojdani E, et al. Infections, toxic chemicals and dietary peptides binding to lymphocyte receptors and tissue enzymes are major instigators of autoimmunity in autism. Int J Immunopathol Pharmacol. 2003;16:189–199.
  39. Gotschall E. Breaking the vicious cycle: intestinal health through diet Kirkton, Ontario, Canada: Kirkton Press; 1994.
  40. Vargas SL, Patrick CC, Ayers GD, et al. Modulating effect of dietary carbohydrate supplementation on Candida albicans colonization and invasion in a neutropenic mouse model. Infect Immun. 1993;61:619–626.
  41. Feingold B. Why your child is hyperactive. New York, NY: Random House; 1985.
  42. Goldstein S, Schwebach A. The comorbidity of Pervasive Developmental Disorder and Attention Deficit Hyperactivity Disorder. J Autism Dev Disord. 2004;34:329–339.
  43. Dengate S. Fed up. North Sydney, Australia: Random House Australia; 2008.
  44. Johnson RR, McClure KE. High fat rations for ruminants. I. The addition of saturated and unsaturated fats to high roughage and high concentrate rations. J Anim Sci. 1972;34:501–509.
  45. Evangeliou A, Vlachonicolis I, Mihailidou H, et al. Application of a ketogenic diet in children with autistic behavior: pilot study. J Child Neurol. 2003;18:113–118.
  46. Ahearn WH, Castine T, Nault K, et al. An assessment of food acceptance in children with autism or pervasive developmental disorder-not otherwise specified. J Autism Dev Disord. 2001;31:505–511.
  47. Ahearn WH. Using simultaneous presentation to increase vegetable consumption in a mildly selective child with autism. J Appl Behav Anal. 2003;36:361–365.
  48. Schreck KA, Williams K. Food preferences and factors influencing food selectivity for children with autism spectrum disorders. Res Dev Disabil. 2006;27:353–363.

CORRESPONDENCE Pramila Srinivasan, PhD 3188 Ross Road Palo Alto, CA 94303 USA E-MAIL: