DESIGN NARRATIVE:

The data were analyzed to estimate the incidence of clinical cardiotoxicity as measured by sudden death, congestive heart failure, or discontinuation of therapy based on cardiac function. Evaluation of patient characteristics (age, anemia) and treatment factors such as drug, dose level, dosing schedule, exposure to irradiation and/or cyclophosphamide identified groups at particularly high risk for development of clinical cardiotoxicity and provided estimates of this risk for future treatment planning. Such estimates of high risk groups should make possible future trials to test the feasibility of using cardioprotectors or alternate dosing schedules to prevent cardiotoxicity. The incidence of clinical cardiotoxicity was calculated using Kaplan- Meier estimates as a function of total cumulative anthracycline dose and also as a function of the time since the end of treatment stratified by dose levels. The estimates were stratified by exposure to cyclophosphamide and radiation therapy. Multivariate methods were used to evaluate the prognostic significance of selected patient characteristics and treatment parameters and to provide estimates of the relative risk of each variable. The method of recursive partitioning was used to identify subpopulations at elevated risk for clinical cardiotoxicity. The data and analytic techniques were accessible through SAS data sets and procedures available to the study at the Pediatric Oncology Group (POG) Statistical Office.