Cell-mediated cytotoxicity is a major effector pathway of the immune system. Thus far, radioactive assays have been widely used, but have significant disadvantages. Meanwhile, flow cytometric assays have been established but have not all been assessed simultaneously relative to the radioactive assays. Here, we have evaluated flow cytometric enumeration of surviving target cells, annexin-V binding and detection of activated caspase-3 and caspase-6 in direct comparison to the 51chromium (51Cr) release assay, and the JAM test. For assay evaluation NKL effector cells Fas-resistant K562 and Fas-sensitive Jurkat target cells were studied. Percent specific lysis measured for each E:T ratio was fitted to a sigmoid dose response curve. Both the flow cytometric and radioactive cytotoxicity assays showed equivalent background lysis (1-13%) but differed considerably with respect to maximum cytotoxicity (11-82% in K562 and 49-75% in Jurkat cells). Half maximum lysis ranged from 4:1 to 28:1 E:T ratios in K562 cells and from 1:3 to 24:1 in Jurkat cells, respectively. Flow cytometric enumeration of surviving target cells was the only assay which permitted detection of cytotoxicity at considerable lower E:T ratios (in K562 cells 1:4 to 2:1 and in Jurkat cells 1:4 to 1:1) than the conventional assays. Prolonged incubation over 24 h did not improve the sensitivity for flow cytometric enumeration of surviving target cells or the JAM test. The observed differences in the lysis of target cells are likely to reflect different sensitivity of cell death-associated changes which are measured by each assay. Thus, the particular choice of a cytotoxicity assay must be carefully adapted to the experimental situation under study.