Cellular expression of the multidrug transporter, P-glycoprotein (Pgp) is recognized as a biological mechanism possibly contributing to treatment failure in acute myeloid leukemia (AML). Correlative studies indicate its association with poor risk features including secondary AML, CD34+ surface phenotype, unfavorable karyotype and advanced age. Reported disparity in the prognostic impact of Pgp relates in part to variance in drug transport capacity. In Pgp expressing cells, capacity for drug extrusion is governed by maturation phenotype and is largely restricted to CD34+ populations lacking myeloid maturation antigens. Three competitive inhibitors of Pgp function showing promise in pilot studies, cyclosporin A (CsA), quinine and the cyclosporin D analogue PSC 833, have entered testing in phase III trials. The presence of non-Pgp-related mechanisms of multidrug resistance, relatively insensitive to Pgp modulators, may limit the success of such treatment strategies. Preliminary investigations indicate that overexpression of the gene encoding the multidrug resistance-related protein (MRP) occurs infrequently in de novo AML, but relative increases in gene message are evident in relapsed specimens. Overexpression of lung resistance protein (LRP) is associated with adverse prognostic variables such as age, secondary disease and Pgp, and has demonstrated prognostic relevance. Because treatment with Pgp modulators may select for this drug resistance phenotype, LRP merits evaluation in randomized trials of Pgp antagonists. These observations indicate that multiple biological mechanisms contribute to anthracycline resistance in AML, thereby warranting development of multifunctional modulators or chemotherapeutic agents with novel mechanisms of action.