One of the possible causes of treatment failure in acute leukemia is the emergence of multidrug resistance caused by P-glycoprotein (P-gp) overexpression. We compared a flow cytometric assay using JC-1 with a technique using rhodamine 123 (rho123) to evaluate the P-gp function in acute leukemia. Samples from 50 acute leukemia patients were analyzed by both functional assays. The P-gp expression was assessed by an immunological flow cytometric test and the association between the P-gp status and the clinical outcome was evaluated. Of all samples, 28% showed a reversible JC-1 efflux and 36% scored positive for the rho123 assay. In two cases, the leukemic blasts showed a reversible JC-1 efflux whereas they were negative for rho123. These patients had blast cells with a very low P-gp activity. Six samples scored positive for the rho123 assay but were negative for the JC-1 test. Five of these samples did not express P-glycoprotein and were considered false positive. We found a strong correlation between the JC-1 and the rho123 test (R(s)=0.59, p<0.0001) and the JC-1 and the immunological assay (R(s)=0.29, P=0.05). There was also an association between the JC-1 status and the clinical outcome of adult patients (chi2=6.30, P=0.04). In conclusion, we recommend the JC-1 assay to study the P-gp activity in acute leukemia because it is more specific and less labor intensive than conventional functional flow cytometric tests using rhodamine 123. In addition, the JC-1 assay can be used to identify adult patients with an increased risk for adverse clinical outcome.