A 4-year prospective study of de novo acute myeloid leukaemia in patients aged 56 years and over was undertaken in the Northern Region of England (population 3.09 million). The study was conducted to assess the incidence and outcome of treatment in all elderly patients diagnosed between January 1, 1988 and December 31, 1991. Two hundred cases de novo AML were confirmed, giving an incidence of 6.05/10(5) per annum (age specific population) (95% Cl, 5.2-6.9). Acute promyelocytic leukaemia was rare. Erythroleukaemia, monocytic leukaemia and AML with trilineage myelodysplasia were more common than in younger patients. Karyotypic abnormalities classically associated with response to therapy were present in only six of 91 patients where cytogenetic data was available. Treatment was at the discretion of the physician in charge: if given, specific treatment was recorded and clinical outcome assessed. Only 84 (42%) of patients received treatment with curative intent. Forty-four of 84 achieved a complete remission, usually of brief duration. A normal karyotype in leukaemic cells was associated with a survival advantage in this group (p < 0.05). Actuarial overall survival at 4 years for the entire group was 2.5%. Even with aggressive treatment, the outcome is poor. The pattern of disease and its lack of response to conventional treatment would support the hypothesis that AML in the elderly may differ biologically from that observed in younger patients. Karyotyping appears to predict those patients likely to benefit from intensive therapy and decisions about management in otherwise fit patients should, if possible, be delayed until a result is obtained. Every effort should be made to give such patients optimal treatment. However, most patients are unsuitable for aggressive treatment and, since long-term survival is rare, cure should not be offered as an inducement to accept such treatment and improving quality of life outside hospital should be the aim of treatment in this group.