This work was designed to discern the frequency of expression of classical multidrug resistance (MDR) in acute myeloid leukemia (AML) at the time of diagnosis, using Western blotting for P-glycoprotein (Pgp) and functional assays for an MDR phenotype (enhancement of daunorubicin [DNR] accumulation/retention and cytotoxicity by the known MDR modulators verapamil, cyclosporin A, and progesterone). Blast cells were studied from 49 newly diagnosed AML patients who were subsequently treated with the "3 and 7" combination of cytosine arabinoside (ara-C) and DNR. DNR accumulation (1 microgram/mL, 3 hours) and retention (16 hours) were determined by flow cytometry. Cyclosporin A (CsA, 5 mumol/L) or verapamil (6.6 mumol/L) each caused significant enhancement of DNR accumulation and retention in these blast cell samples (P < .001, Wilcoxon's test). Verapamil or CsA caused greater than 20% enhancement of DNR accumulation or retention in over 25% or 50% of these patients, respectively; however, there was no correlation with the presence or degree of enhancement and response to treatment. Progesterone (10 mumol/L) caused no significant enhancement of DNR accumulation or retention. The effects of the MDR modulators on the cytotoxicity of DNR was also determined in blast cells from 40 of the patients, using a flow cytometric assay. CsA alone was cytotoxic (caused an approximate 20% decrease in cell survival compared with control, P < .001); CsA or verapamil caused enhancement of 1 mumol/L DNR cytotoxicity (P < .001). Greater than 40% enhancement of cell kill by CsA or verapamil was observed in over 75% of patients studied. There was no difference in the degree of enhancement of cytotoxicity between patients clinically sensitive or resistant to treatment. Progesterone caused no enhancement in DNR cytotoxicity. In contrast to the functional assays, highly sensitive immunoblots using the C219 antibody to Pgp showed evidence of low level expression of Pgp in blast cells from only 3 of these patients: 1 was chemotherapy resistant, 2 were sensitive. Thus, although the functional assays suggest a high frequency of expression of a classic MDR phenotype in AML patients at the time of diagnosis, with enhancement by CsA obtained at a clinically relevant concentration (5 mumol/L), the frequency of Pgp expression detectable by C219 Western blots was low in these patients. This could be interpreted either that the method used was not sufficiently sensitive to detect Pgp in all of the blast cell specimens that actually overexpressed mdr1, or that the accumulation/efflux-based MDR phenotype observed is not always mediated by Pgp in these previously untreated patients.