Malignancies in childhood occur with an incidence of 13-14 per 100,000 children under the age of 15 years. Acute lymphoblastic leukaemia with an incidence of 29% is the most common paediatric malignancy, whereas acute myeloid leukaemias account for about 5%. The treatment of acute leukaemias consists of sequential therapy cycles (induction, consolidation, intensification, maintenance therapy) with different cytostatic drugs over a time period of up to 1.5-3 years. Over the last 25 years of clinical trials, a significant rise in the rate of complete remissions as well as an increase in long-term survival has been achieved. Therefore, growing attention is now focused on the long-term effects of antileukaemic treatment. Several cytostatic drugs administered in the treatment of acute leukaemia in childhood are known to cause long-term adverse effects. Anthracyclines may induce chronic cardiotoxicity, alkylating agents are likely to cause gonadal damage and secondary malignancies and the use of glucocorticoids may cause osteonecrosis. Most of the long-term adverse effects have not been analysed systematically. Approaches to minimising long-term adverse effects without jeopardising outcome have included: the design of new drugs such as a liposomal formulation of anthracyclines, the development of anthracycline-derivates with lower toxicity, the development of cardioprotective agents or, more recently, the use of targeted therapy;alternative administration schedules like continuous infusion or timed sequential therapy; and risk group stratification by the monitoring of minimal residual disease. Several attempts have been made to minimise the cardiotoxicity of anthracyclines: decreasing concentrations delivered to the myocardium by either prolonging infusion time or using liposomal formulated anthracyclines or less cardiotoxic analogues, or the additional administration of cardioprotective agents. The advantage of these approaches is still controversial, but there are ongoing clinical trials to evaluate the long-term effects. The use of new diagnostic methods, such as diagnosis of minimal residual disease, which allow reduction or optimisation of dose, offer potential advantages compared with conventional treatment in terms of reducing the risk of severe long-term adverse effects. Most options for minimising long-term adverse effects have resulted from theoretical models and in vitro studies, but only some of the modalities such as the use of dexrazoxane, the continuous infusion of anthracyclines or timed sequential therapy, have been evaluated in prospective, randomised studies in patients. Future approaches to predict severe toxicity may be based upon pharmacogenetics and gene profiling.