A pilot prospective follow-up study of the role of the ketogenic diet was carried out on 30 children, aged between 4 and 10 years, with autistic behavior. The diet was applied for 6 months, with continuous administration for 4 weeks, interrupted by 2-week diet-free intervals. Seven patients could not tolerate the diet, whereas five other patients adhered to the diet for 1 to 2 months and then discontinued it. Of the remaining group who adhered to the diet, 18 of 30 children (60%), improvement was recorded in several parameters and in accordance with the Childhood Autism Rating Scale. Significant improvement (> 12 units of the Childhood Autism Rating Scale) was recorded in two patients (pre-Scale: 35.00 +/- 1.41[mean +/- SD]), average improvement (> 8-12 units) in eight patients (pre-Scale: 41.88 +/- 3.14[mean +/- SD]), and minor improvement (2-8 units) in eight patients (pre-Scale: 45.25 +/- 2.76 [mean +/- SD]). Although these data are very preliminary, there is some evidence that the ketogenic diet may be used in autistic behavior as an additional or alternative therapy.