P-glycoprotein (Pgp) is a trans-membraneous protein that is associated with multidrug resistance (MDR) in human cancer, including hepatocellular carcinomas and leukemias. There is no consensus regarding methods of choice for analysis of Pgp expression, and development of reliable analytical methods is now essential. We have studied the the Pgp expression in human hepatoma and leukemia cell lines using flow cytometry. The aim of the study was to compare binding properties of anti-Pgp antibodies reacting with surface (MRK16, UIC2) and cytoplasmic (C219, JSB-1) epitopes to assess which antibody performed best with respect to fluorescence discrimination. By histogram subtraction the fractions of resistant human hepatoma cells positive for Pgp were 99% (MRK16), 97% (UIC2), 77% (JSB-1), and 51% (C219), demonstrating variations in antibody reactivity. The resolution in detecting decreasing levels of Pgp in hepatoma cells was superior for the externally binding antibodies, showing that there is a correlation between antibody reactivity and fluorescence discrimination. Similar results were obtained for parental and resistant KG1a human leukemia cell lines. The Pgp epitopes remained reactive to the anti-Pgp MAbs after methanol fixation and cryopreservation. By dual parameter flow cytometry it was shown that Pgp expression in viable cells may be assessed together with uptake of epirubicin, which was low in cells expressing high levels of Pgp and vice versa. In conclusion, all tested antibodies proved useful for flow cytometric detection of high levels of Pgp, but the externally binding ones were superior in detection of low and variable levels of Pgp.