Very few statements can be made about the epidemiology, etiology, course, or treatment of childhood autism that are not arguable, including this one. As a contribution to a series in Clinics in Developmental Medicine, this update of a 1985 book focusing on neurobiological aspects by a Swedish (Gothenberg) professor of child and adolescent psychiatry and an American (Georgetown) professor emerita of pediatrics provides a provocative compendium motivated by a desire to replace much received wisdom with their own brand of received wisdom. Both of the editions represent milestones on the pathway to knowledge about autism that mark a sharp divergence from beliefs launched by Leo Kanner in 1943 when he reported the first l l cases of this perplexing phenotype and led-read misled-parents and clinicians into concluding that the syndrome was caused by the refrigerator-like personalities of the parents. He later (1969) recanted his belief popularized by Time magazine. The upward social class bias of the earliest referrals led to another myth, later debunked by proper epidemiological studies including those by the authors, that the condition was limited to the offspring of professional, upperclass parents.
Autism is classified in ICD-I0 on Axis I1 as the first of six Pervasive Developmental Disorders (F84.0) [nominally listed as a probable autosomal recessive in McKusick, 1992, but with a commentary leaving all alternatives open], together with Atypical autism, Rett syndrome (F84.2) [a probable X-linked dominant in McKusick, 19921, Other disintegrative disorders, Overactive disorders associated with mental retardation and stereotyped movements, and Asperger's syndrome (F84.5) [schizoid disorders of childhood]. Narrow classification of the phenotype requires three areas of dysfunction to become manifest by age 3 years-communication deficits (including no phrases by age 33 months as one clear example), impairments in reciprocal social interaction (caregivers or peers), and repetitive stereotyped patterns of interests/activities. Such rules result in a prevalence of childhood autism of 4 per 10,000 with a sex ratio of 3-4 males to each female. Some 65 to 85% of cases have IQs less than 70, and mental retardation per se is not an exclusion criterion, nor is a concomitant medical condition. In a valuable tabling of 14 epidemiological studies conducted through 1991, 6 involving Gillberg himself, it can be seen that a broadening of the