All-trans retinoic acid (ATRA) is a known inducer of differentiation in acute promyelocytic leukemia. To improve the outcome of children with acute promyelocytic leukemia, ATRA has been applied since 1994 as an additional induction element inthe AML-BFM 93 study. In a retrospective study, we compared 22 children treated with ATRA (median age: 9.3 years; range: 1.8-16.3) with 22 patients receiving conventional therapy (median age: 12.3 years; range: 3.2-16.7). Twenty-one of the children achieved complete remission. Only one patient died early from bleeding complications after 3 days administration of ATRA. In the control group, seven early deaths occurred (Fisher exact test; p<0.04). Two children died from intracerebral hemorrhages. Two patients suffered from sepsis during aplasia after induction therapy, and one child did not respond to treatment. The 5-year overall survival (OS) and event-free survival (EFS) of the children who received ATRA followed by chemotherapy were significantly bettercom-pared with conventionally treated children [OS: 0.87 +/- 0.9 vs 0.45 +/- 0.11, p (log rank) <0.003; EFS: 0.76 +/- 0.11 vs 0.43 +/- 0.11 p (log rank) <0.02]; the median observation time was 2.8 years (19-76 months). However, nearly all children suffered from common side effects such as headache, fever, joint, muscle and bone pain, weight gain, or dermatitis. In three patients, a retinoic acid syndrome was observed. Interruption of ATRA treatment and application of dexamethasone, necessary in 12 children, controlled theadverse effects. ATRA treatment could be resumed in 18 patients. In conclusion, ATRA treatment during induction could avoid early deaths in children with acute promyelocytic leukemia with considerable but manageable toxic side effects.