This study will assess the effectiveness of the ketogenic diet (high-fat, low-carbohydrate, and moderate protein) in treating autism spectrum disorder (ASD). Three study groups will be comprised of children (2-21 years of age) based on whether or not they have ASD and receive the ketogenic diet - ASD/ketogenic diet, ASD/non-ketogenic diet, and non-ASD/non-ketogenic diet.

Recent studies have shown that the ketogenic diet (high-fat, low-carbohydrate, and moderate protein; induce a shift from the primary metabolism of glucose to ketones) or Modified-Atkins diet may be effective in treating autism. Research on the Black and Tan BRachyury (BTBR) T+tf/J mouse strain, characterized by an autism-like behavioral phenotype, has demonstrated the efficacy of a ketogenic diet in improving autism. Although, modified diets, such as the Feingold diet, low-sugar diet, or gluten-free diet, have shown behavioral improvements in patients with Attention Deficit Hyperactivity Disorder (ADHD), the ketogenic diet has not been studied in autism spectrum disorder (ASD) despite the high incidence of ADHD comorbidity. The findings from only one prospective, pilot study have been published, which reported significant behavioral improvement in all 18 autistic subjects after six months (assessed at intervals of four weeks on the diet and two weeks diet-free) on the ketogenic diet. Clearly, the ketogenic and other modified diets as promising treatments for ASD have been understudied. Additional clinical research is necessary to establish the ketogenic diet as a safe, effective treatment alternative for children with ASD.

This proposed research project will involve an ethnically diverse sample (varied genetic background and environmental exposure) to assess the efficacy of the ketogenic diet as a therapeutic intervention, and to understand its beneficial effects in children with ASD. The investigators anticipate that noteworthy findings will contribute to the sparse literature on ASD and effective dietary interventions and prompt future research collaborations with investigators from other medical centers and/or academic institutions. Funding for future research appears promising considering that ketogenic therapies are also a novel method for the treatment of a variety of disorders, including other neurodevelopmental disorders, diabetes, migraine headaches, brain tumors, multiple sclerosis, and obesity.

In addition, the investigators propose to also examine the potential changes in blood composition and intestinal (gut) microbiome in children with (without) ASD who are treated (not treated) with the ketogenic diet in light of the effect on their ASD symptomatology. The relationship between the gastrointestinal tract and the central nervous system has garnered increasing interest in the research community. The gut-brain interface describes a bidirectional relationship in which the central nervous system influences the gut and vice versa. Little is known about the mechanisms behind anecdotal reports of dietary success in humans, but it is suspected that alterations in gut microflora are involved. Mulle et al. (2013) postulated that the connection between the gut microbiome and ASD may be either the direct cause or as indirect consequence of atypical patterns of feeding and nutrition. Similarly, diet patterns, including the ketogenic diet, impact nutrient availability and thus indirectly modulate the gut microbiome. Multiple studies have confirmed differences in levels of gut commensals and overall metabolite profiles in fecal and urinary samples from healthy children compared to children with ASD, potentially as a result of changes in the microbiome. In the maternal immune activation mouse model of ASD, Hsiao et al. (2013) reported that probiotic treatment with B. fragilis could correct behavioral abnormalities and metabolomics profile typified in ASD and ameliorate ASD-relevant GI barrier deficits in mice. Sandler et al. (2000) demonstrated that oral vancomycin treatment showed short-term improvements in children with regressive-onset autism, suggesting that alterations in gut bacteria profile, in this case with antibiotic therapy, may improve autistic behavior. Studies of the gut microbiome in children with ASD may elucidate the role of diet and the alterations in gastrointestinal microbes related to ASD. As a result of these data, novel treatments may be discovered.

The specific aims of this study are:

Aim 1. To evaluate the effect of the ketogenic diet on the core symptoms of autism. The following instruments will be used to measure core autistic symptoms: a) Autism Diagnostic Observation Schedule - Second Edition (ADOS-2); b) Asperger Syndrome Diagnostic Scale (ASDS); c) Childhood Autism Rating Scale (CARS-2); d) Gilliam Autism Rating Scale (GARS-3); e) Social Responsiveness Scale - Second Edition (SRS-2); f) Diagnostic and Statistical Manual IV Text Revision (DSM-IV-TR) and DSM-V ASD criteria; g) Standardized social & intelligence tests (if available, administered by child's school); and h) Vanderbilt ADHD Diagnostic Teacher Rating Scale Forms (Vanderbilt).

Hypothesis: Participants who have ASD/on the ketogenic diet will have significantly lessened core autistic symptoms than those that have ASD/not on the ketogenic diet, between baseline to three and six months after the dietary intervention is initiated.

Aim 2. In anticipation of significant changes in core autistic symptomatology, to examine the effects of the ketogenic diet on the (a) number and dosage of medications used for behavioral management, (b) number of lab tests ordered for behavioral management, (c) number of emergency room or hospital visits for behavioral management, and (d) subject/family satisfaction with the ketogenic diet.

Hypothesis: The number (and/or dosage) of medications, lab tests ordered, and emergency room or hospital visits for behavioral management will decrease, and participant/family satisfaction will be high for participants who have ASD/on the ketogenic diet than those who have ASD/not on the ketogenic diet, between baseline to three and six months after the dietary intervention is initiated.

Aim 3. To compare differences and/or changes in (a) biochemical profiles as defined from blood and stool (gut or fecal microbiome) specimen samples and (b) serum and urine ketone levels.

Hypothesis: Participants who have ASD/on the ketogenic diet will have notably different biochemical profiles and significantly higher serum/urine ketone levels than those who have ASD/not on the ketogenic diet and typically developing controls on a regular diet, between baseline to three and six months after the dietary intervention is initiated.

The investigators anticipate that the KD will be effective in treating ASD, and further an understanding of dietary patterns and gastrointestinal alterations in ASD. Noteworthy findings will contribute to the sparse literature on the association between dietary intervention and neurodevelopmental disorders, and assist with obtaining future funding for higher-level clinical trials involving collaborative research with other medical centers and academic institutions. These initiatives, the investigators believe, are necessary to establish the KD as a novel, safe alternative to effectively treat patients with ASD.

Inclusion Criteria:

• Ages 2-21 years.
• Primary diagnosis of autism spectrum disorder.
• Parent/legal guardian and child able to read or understand English, and able/willing to provide informed consent/assent.
• Females of childbearing potential must have a negative pregnancy test result and agree to use a medically acceptable method of contraception throughout the entire study period and for 30 days after the last dose of study drug - childbearing potential is defined a girls who are > Tanner stage 2 and urine pregnancy tests are acceptable.

Exclusion Criteria:

• Known cardiac disorder including arrhythmias or hypertension.
• BMI < 3rd%ile.
• Carnitine deficiency (primary).
• Carnitine palmitoyltransferase (CPT) I or II deficiency.
• Carnitine translocase deficiency.
• Beta-oxidation defects - medium-chain acyl dehydrogenase deficiency (MCAD), long-chain acyl dehydrogenase deficiency (LCAD), short-chain acyld dehydrogenase deficiency (SCAD), long-chain 3-hydroxyacyl-coenzyme A (CoA) deficiency, and medium-chain 3-hydroxyacyl-CoA deficiency.
• Pyruvate carboxylase deficiency.
• Porphyria.
• Inability to maintain adequate nutrition.
• Patient or caregiver non-compliance.