Maintenance treatment for AML is an approach to minimize residual disease, optimize quality of remission and prevent a leukemic regrowth over a longer period of time. This intention implies a certain antileukemic activity and myelotoxicity. Thus, a prolonged myelosuppressive maintenance is best exemplified by the optimized protocol of the CALGB published by Kanti R. Rai in 1981 (Blood 58:1203-1212, 1981) and derived by the AMLCG as a therapeutic standard. From our today's knowledge about the impact of various strategies, a lack of postremission therapy is not compatible with durable remissions. Even after an induction-type consolidation, the classic CALGB-type maintenance, or a comparably intensive regimen improved the relapse-free survival over that from alternatives. Some studies which failed to show a benefit used maintenance at low-dosage or short duration. Data about maintenance delivery in patients reaching long-term remissions demonstrate feasibility and compliance, and a low maintenance-related death rate can compete with that from alternative options. Revisiting maintenance, however, requires a comparison with other strategies on the basis of intention-to-treat. Either single prospective trials or crosstrial networking by a common standard arm and general upfront randomization can further assess the relative value of maintenance for AML.