This naturalistic, cross-sectional study was designed to assess the risk of prolactin level elevation and associated side effects in youths taking long-term atypical antipsychotic medication. Subjects were enrolled from outpatient child psychiatric treatment settings in upstate New York who were taking risperidone, olanzapine, or quetiapine for at least 6 months. Demographic data, medication history, and side effects were elicited at the initial interview. Two fasting morning serum prolactin levels were obtained 1 month apart, and the results were averaged. Fifty outpatient youths, with a median age of 13 years, were enrolled in the study. The median overall duration of use of an atypical antipsychotic was 22.1 months. The median dose of medication for risperidone was 1.5 mg/day, for olanzapine 10 mg/day, and for quetiapine 200 mg/day. The mean prolactin level among all patients on risperidone was significantly greater than controls, as well as for those on quetiapine or olanzapine. The risk of hyperprolactinemia with long-term use of risperidone appears to be significantly greater than for olanzapine or quetiapine. Overt side effects were infrequent in the overall sample, but serum prolactin assessment is recommended for youths taking risperidone chronically. Because of variability found in sequential prolactin samples, repeat samples may be warranted.