OBJECTIVE: The objective of this study was to assess the prognostic value of response to prednisone in children with acute lymphoblastic leukemia (ALL). METHODS: One hundred and twenty-six children with newly diagnosed ALL were enrolled in the protocol of ALL-XH-90. A prednisone good response (PGR) was defined as a blast count of less than 1 000/μL and a prednisone poor response (PPR) as a blast count of at least 1 000/μL in peripheral smears, after 7 days of oral prednisone (60 mg/m~2) and one intrathecal injection of methotrexate along with cytosine arabinoside. The probability of event-free survival (pEFS) was estimated by Kaplan-Meier analysis and the differences in pEFS between groups were assessed by the log-rank test. Differences in the biologic characteristics between groups were compared by Chi-square analysis or Fisher exact test (two-tail). RESULTS: One hundred and ten patients had PGR and 16 had PPR. The 5-year pEFS was significantly worse for patients with PPR compared with those with PGR (48%±13% vs 73%±5%, P= 0.0021). CONCLUSIONS: Response to prednisone is a simple, practical and prognostic factor in childhood ALL.