Objective. To determine the early and late cumulative incidence of anthracycline-induced clinical heart failure (A-CHF) after anthracycline therapy in childhood and to identify associated risk factors. Design. Retrospective study of the case records. Patients and methods. Data were derived from the case histories of the 607 children who, according to the records of the Childhood Tumours Working Group, had been treated with anthracycline derivatives in the period from January Ist, I976 to December 3Ist, I995 in the Emma Children's Hospital of the Academic Medical Centre in Amsterdam. The cumulative incidence of clinical heart failure was calculated from these data by the Kaplan-Meier method and the risk factors for heart failure were determined by multivariate regression analysis. Results. The clinical status up to at least December 3Ist, I996 could be determined for 580 children. The cumulative incidence of A-CHF was 2.8%, after a mean follow-up time of 6.3 years and a mean cumulative dose of anthracycline derivatives of 30I mg/m 2 . A cumulative dose of anthracycline higher than 300 mg/m 2 was associated with an increased risk of A-CHF (relative risk: II.8; 95% CI: I.6-59.5) compared with a cumulative dose lower than 300 mg/m 2 . The estimated risk of A-CHF increased with time after the start of anthracycline chemotherapy: from 2% after 2 years to 5% after 15 years. Conclusion. Up to 5% of the children will develop A-CHF I5 years after treatment with anthracycline derivatives; patients treated with a cumulative dose higher than 300 mg/m 2 are at greatest risk for A-CHF.