The present article reports a multiparametric cytofluorimetric analysis of apoptosis in murine thymocytes and human PBMC from healthy donors or HIV-infected patients. We have evaluated four previously described cytofluorimetric methods of apoptosis quantification, each of them detecting distinct cellular alterations of the apoptosis process. Reduced DNA stainability was detected with the PI assay on nuclei and the AO/EB dual staining method was evaluated on entire and non-fixed cells. DNA strand breaks were detected following in situ nick translation, and alterations in membrane integrity were evaluated following 7-AAD incorporation. When apoptosis was quantified in murine thymocytes under various conditions of induction, the combined analysis of FSC/SSC criteria and 7-AAD or AO/EB staining on the same samples permitted the identification of distinct steps in the apoptosis process. Moreover these four methods proved to be reliable and gave statistically similar results both on murine thymocytes and PBMC from healthy donors. However, in HIV-infected persons, some discordant apoptosis determinations were observed with PI and 7-AAD staining assays. We found that after Ficoll isolation, PBMC from AIDS patients were enriched in erythrocytes and granulocytes. On the one hand, granulocytes were found to be responsible for a poor apoptosis estimation with the PI assay whereas erythrocytes were responsible for an underestimation rate of apoptosis in the 7-AAD assay. To prevent such interference, we propose some modifications which render these methods more suitable for application to PBMC from HIV-infected patients. Taken together these observations indicate that it is essential to assess critically the apoptosis quantification methods with respect to their applicability to complex lymphoid populations such as those from AIDS patients.