Parents who research treatments for autism are confronted with a bewildering array of options, almost all of which have never been tested for safety and effectiveness. Organizations like The Cochrane Collaboration, which reviews the quality of evidence for medical treatments, are putting more effort into evaluating popular alternative treatments.

So far, the most comprehensive review of alternative autism treatments comes from two pediatricians: Susan Hyman of the University of Rochester School of Medicine Golisano Children's Hospital at Strong and Susan Levy, a clinical professor of pediatrics at the University of Pennsylvania School of Medicine and The Children's Hospital of Philadelphia. Their 2008 analysis gave each treatment a letter grade for the quality of the research conducted up to that point; the mark, however, is not a ranking of the treatment's safety or effectiveness.

The two pediatricians based the grades on the amount of testing done on the treatments, which in most cases was skimpy at best. Research that got an "A" grade included randomized control trials, the gold standard for medical research, and meta-analyses, which compare research from different labs. A "B" went to treatments that had been studied in "well-designed controlled and uncontrolled trials," according to Hyman. The "C" grades, the lowest category (there were no "D"s or "F"s), were based on case reports, theories and anecdotes, which are not considered acceptable for mainstream medical research.

Research on just one treatment, secretin, was good enough to earn an A. In short, there is a lot more work that needs to be done toward testing popular alternative treatments and getting more potential treatments into development at research institutions and pharmaceutical companies.

Dietary supplements

B6/Mg++—Grade: B

Vitamin B6 and magnesium have been a popular treatment for autism over the past 20 years. The Cochrane Review identified three studies that compared outcomes of B6 and magnesium treatment with those for placebo or no treatment, but just 28 subjects were treated altogether. One study found no improvements; another reported improvement in IQ and social behaviors. But all the studies suffered methodological weaknesses aside from the small sample size.

DMG—Grade: B

Dimethylglycine (DMG), an antioxidant and derivative of the amino acid glycine, is marketed as an immune system booster. Two small double-blind studies of DMG found it had no effect on autism symptoms.

Melatonin—Grade: B

Melatonin is a hormone produced by the pineal gland that regulates sleep. Melatonin supplements are popular for self-treating insomnia or jet lag. Many people with autism-spectrum disorders report sleeping problems, and at least one study has found improvements in falling asleep and staying asleep.

Vitamin C—Grade: B

Vitamin C, an antioxidant, is often part of vitamin supplements given to children with autism. One study reported less repetitive behavior in a double-blind, placebo-controlled trial of vitamin C in 18 children with autism.

Amino Acids—Grade: C; L-Carnosine—Grade: B

Neurotransmitter abnormalities have long been a focus of autism research. Some amino acids act as neurotransmitters or prompt their production, so amino acids like tryptophan have been tried as alternative treatments. No trials have studied the benefits of supplementation with tryptophan, taurine, lysine or GABA. L-carnosine, a molecule made of two amino acids that has antioxidant properties, is marketed as an anti-aging remedy. One double-blind, placebo-controlled trial of L-carnosine in 31 children with autism found improved expressive and receptive vocabulary.

Omega-3 fatty acids—Grade: B

Polyunsaturated fatty acids, in particular omega-3 fatty acids, are crucial for brain development and cannot be manufactured in the body. Essential fatty acid supplements such as fish oil have become popular for children with autism. A recent randomized, double-blind, placebo-controlled, six-week pilot study found behavior improvements in 13 children with severe behavior problems as a result of autism.

Folic acid—Grade: C

Oxidative stress is a theory that some people have advanced to account for the atypical brain development seen in autism, and abnormal levels of antioxidants have been reported in children with autism. But there are no randomized, controlled trials testing the notion that supplementation with folic acid, a water-soluble B vitamin that helps produce and maintain new cells, would have beneficial effects.

Secretin—Grade: A

Secretin, a gastrointestinal hormone, is one of the most extensively studied autism treatments. More than a dozen well-designed, well-executed studies have failed to find any benefit.

Pharmaceutical treatments

Antibiotics—Grade: C

Parent reports of frequent respiratory or gastrointestinal infections in children with autism are used to support the theory that the children have immune system problems, but those findings have not been confirmed. One study found short-term behavioral improvement in 11 children treated with oral vancomycin. But there are no other data supporting the use of antibiotics, and the researchers in that study said they would not recommend it for routine treatment.

Antifungal agents—Grade: C

Treatment with antifungal agents is based on the premise that imbalances in intestinal flora or other immune factors lead to an overgrowth of yeast. No controlled trials have tested antifungals as an autism treatment despite the popularity of medications such as nystatin (Mycostatin) and fluconazole (Diflucan).

Gastrointestinal medications—Grade: C

Children with autism frequently have symptoms such as reflux, constipation and diarrhea, and eat only a very limited number of foods. There are no evidence-based studies on the efficacy of digestive enzymes or probiotics for treating these symptoms.

Hyperbaric Oxygen Therapy—Grade: C

Hyperbaric oxygen therapy (HBOT) is used in conventional medicine to treat carbon monoxide poisoning and to speed wound healing, and it has become popular as an autism treatment based on theories that implicate gut or brain inflammation or lack of blood flow to the brain. There are no randomized clinical trials of HBOT for autism. One open trial of 18 children with autism found some decrease in C-reactive protein, a marker for inflammation, and parents reported improved behavior. But the subjective measures and the fact that many of the children were also taking antioxidant supplements "make this study difficult to interpret," Hyman and Levy report.

Immune therapies—Grade: C

Some alternative practitioners recommend treatment with intravenous immunoglobulin-G, on the premise that immune deficits cause symptoms of autism. One open trial reported subjective improvements, but two other trials with specific outcome measures found no benefit.

Other approaches

Chelation—Grade: C

One alternative theory holds that mercury is poorly eliminated by children with autism, and that the toxic metal alters immune function and development. Epidemiological studies have failed to find a link between the use of the ethyl mercury–based preservative thimerosal in vaccines and autism. Despite this, chelation, the standard treatment for heavy metal poisoning, is marketed as an off-label treatment for autism. There are no controlled studies testing chelation's safety or effectiveness as an autism treatment, and at least one child has died after being treated with EDTA (ethylenediamine tetraacetic acid) chelation for autism.

Gluten-free/casein-free diet (GF/CF)—Grade: B

This diet is a popular alternative treatment for autism, based on the premise that the proteins gluten (found in wheat) and casein (in milk) aggravate autistic symptoms because they in some way mimic opiate neuropeptides. One small single-blind trial found some improvements; larger double-blinded trials have found none. Hyman and Levy speculate that improvements seen by parents may be from removing lactose from the diets of children who are lactose intolerant.

Source: Levy, S.E. & Hyman, S.L. Child Adolesc. Psychiatr. Clin.  N. Amer. 17, 803-820 (2008).
Child Adolesc Psychiatr Clin N Am. 2008 October; 17(4): 803–ix.