Recent research has yielded increasing support for neurobiologic theories of autism. A number of family and twin studies support the role of genetics and have led to wide acceptance of autism as an organically based disorder. Controversy persists, however, over the role of congenital medical conditions in the etiology of autism. Two rather divergent views have emerged. One, advocated by Gillberg and colleagues, proposes that up to 30% of cases of autism are associated with a known medical condition. On the other hand, research by Rutter and colleagues suggests the incidence may be closer to 10%. In this retrospective study records on 211 subjects with autism and other developmental disorders are reviewed to determine the prevalence of associated medical conditions and its variability related to the system used to diagnose autism. Results suggest the prevalence of medical conditions with suspected etiologic relationship with autism varies between 10 and 15%, depending on the diagnostic system employed. Further variability in prevalence rates results from a less strict definition of "medical condition" and yields rates between 25 and 37%. Disparate findings in previous research may stem from variability in both diagnostic system employed and which medical conditions are considered significant in the etiology of autism.