Factors related to haloperidol response and dyskinesias in autistic children.

A secondary analysis of data pooled from three studies (Anderson et al. 1984, 1989; Campbell et al. 1978) was performed to identify variables predictive of haloperidol response in 125 autistic children, with ages ranging from 2.3 to 8.2 years. Mean behavioral improvement was greater under haloperidol treatment conditions than under placebo. Higher intelligence quotient (IQ) was predictive of reduction in behavioral symptoms under general conditions of haloperidol or placebo treatment, while older children were found to respond favorably to haloperidol itself. Under both haloperidol and placebo conditions, there was also a tendency for greater reduction in symptoms, in terms of raw score and percent change, for those with greater initial severity of illness. Results for initial severity of illness as a predictor of improvement generalized across a wide variety of behavior not specific to autism (e.g., hyperactivity and temper outbursts). However, mean behavioral improvement and its prediction with demographics for individuals tended to be more specific to symptoms related to autism per se. Reduction in symptoms during short-term haloperidol treatment was not found to be related to whether or not children developed dyskinesias in subsequent long-term haloperidol administration.