OBJECTIVE To determine the impact of treatment toxicity on long-term survival in pediatric Hodgkin's disease. METHODS We studied late events in 387 patients treated for pediatric Hodgkin's disease on four consecutive clinical trials at St Jude Children's Research Hospital from 1968 to 1990. Relative risks, actuarial risks, and absolute excess risks for cause-specific deaths were calculated. RESULTS As of April 1997, 316 (82%) of patients were alive, with a median follow-up of 15.1 (range, 2.9 to 28.6) years. In this cohort, which represented 5,623 person-years of follow-up, 71 fatal events resulted from Hodgkin's disease (n=36), second malignancies (n=14), infections (n=7), accidents (n=7), cardiac disease (n=6), and asphyxiation (n=1). The 5-year estimated event-free survival (EFS) for the entire cohort was 79.6%+/-2.1 %, which declined to 63.1%+/-4.4% by 20 years. Cumulative incidences of cause-specific deaths at 25 years were 9.8%+/-1.6% for Hodgkin's disease, 8.1%+/-2.6% for second malignancy, 4.0%+/-1.8% for cardiac disease, 3.9%+/-1.5% for infection, and 2.1%+/-0.8% for accidents. Standardized incidence ratios showed excess risk for all second malignancies (12; 95% confidence interval [CI], 8 to 17), acute myeloid leukemia (81; 95% CI, 16 to 237), solid tumors (11; 95% CI, 7 to 16), and breast cancer (33; 95% CI, 12 to 72). Standardized mortality ratios also showed excess mortality from cardiac disease (22; 95% CI, 8 to 48) and infection (18; 95% CI, 7 to 38). CONCLUSIONS Compared with age- and sex-matched control populations, survivors of pediatric Hodgkin's disease who were treated before 1990 face an increased risk of early mortality related to second cancers, cardiac disease, and infection.