The records of twenty-six infants with both symptomatic infantile spasms and classic hypsarrhythmia were reviewed to determine the efficacy of various ACTH dosages and time of initiation of therapy. Mean age of infantile spasm onset was 6.4 months. Most patients (13) had sustained perinatal hypoxic-ischemic insults. Seventeen patients (65%) had complete cessation of spasms. Between these responders and the 9 nonresponders there was no difference in duration of spasms prior to treatment (2.6 and 2.0 months) or mean ACTH dose (87.4 and 84.5 U/m2, respectively). Infants treated with high-dose ACTH (> 100 U/m2) did not have an improved response rate. The most favorable outcomes were associated with spasm onset at > 8 months of age (all of whom were responders, regardless of dose) or when treatment was started within 1 month of onset of infantile spasms with > 80 U/m2 ACTH (88% responders). Infants treated more than 2 months after onset often did not respond (57%) regardless of dose. Nonresponders with spasm onset at < 4 months of age had the worst prognoses; all had poorly controlled seizures and regressed developmentally. Although all infants in the study were neurologically abnormal, development either improved or did not deteriorate in most responder infants following spasm resolution and one-half remained seizure free. Nonresponder infants continued to have infantile spasms or other seizure types. These data suggest that ACTH is valuable in the treatment of significantly impaired infants with symptomatic infantile spasms, but the most important determinants of outcome may be age of onset and rapidity of treatment rather than dosage.