To determine the incidence of and risk factors for the development of secondary acute leukaemia and myelodysplasia in patients treated in British National Lymphoma Investigation's studies of Hodgkin's disease since 1970. 2676 Patients entered into Hodgkin's disease studies between February 1970 and November 1986. Data accrued up to November 1988 were analysed, ensuring a minimum follow up period of two years. Retrospective analysis of multicentre trial data by case-control and life table methods. 17 Cases of secondary leukaemia were recorded in this group of 2676 patients, giving an overall risk at 15 years of 1.7%. The risks of leukaemia after chemotherapy alone and chemotherapy with radiotherapy were not significantly different. The risk of leukaemia increased sharply with the amount of treatment given as measured by the number of attempts at treatment. The 15 year risks of leukaemia were 0.2%, 2.3%, and 8.1% for patients receiving one, two, or three or more attempts at treatment. The highest risk, 22.8% at 15 years, was observed in patients treated with lomustine (CCNU), and a case-control study suggested that this was an independent risk factor. The risk of secondary leukaemia was largely related to the overall quantity of treatment, although exposure to lomustine seemed to be an important risk factor. Treatment with both drugs and radiation was not more leukaemogenic than treatment with drugs alone. The greatest risk of secondary leukaemia was seen in multiply treated patients who were unlikely to be cured of Hodgkin's disease. Avoidance of secondary leukaemia should be a minor factor in the choice of treatment for Hodgkin's disease.