The flow cytometric enumeration of CD34+ hemopoietic precursor cells (HPC) present in samples used for transplantation of HPC has proven to be the most powerful single parameter for prediction of engraftment. At present, several different methodological approaches are used for the flow cytometric enumeration of CD34+ HPC. In the present study we have compared two of these methods as regards enumeration of CD34+ HPC and their CD34+/CD19- and CD34+/CD19+ subsets: a lyse-non-wash procedure based on the use of a recently commercialized red cell lysing solution (Quicklysis, Cytognos, Salamanca, Spain) and a lyse-and-then-wash method in which the Becton Dickinson (San Jose, CA) FACS Lysing Solution was used. For that purpose a total of 52 samples corresponding to 20 G-CSF mobilized peripheral blood (PB) samples and 21 PB-derived leucapheresis products from patients undergoing autologous PB stem cell harvest, together with 11 bone marrow (BM) samples from healthy volunteers were analyzed. Our results show that for each of the three types of samples analyzed the use of the lyse-and-then-wash method is associated with significantly lower numbers of both total CD34+ HPC (P < or = 0.003) and its major CD34+/CD19- subset (P < or = 0.01) while no significant changes are detected in the number of CD34+/CD19+ HPC in BM samples (P > 0.05). The use of an internal standard (reference beads) added just prior to data acquisition, showed that the differences between both methods are due to a selective loss of CD34+ HPC and its major CD34+/CD19- subset in BM (P=0.002 and P=0.003), PB (P < 0.0001 and P < 0.0001) and PB-derived leucapheresis products (P < 0.0001 and P=0.0001). Finally, addition of a centrifugation and washing step to a group of 11 leucapheresis samples lysed with Quicklysis showed that they did not significantly affect the overall number of total CD34+, CD34+/CD19- and CD34+/CD19+ HPC obtained. In line with these findings elimination of centrifugation and washing steps when FACS Lysing Solution was used to lyse mature red cells almost corrected for the selective loss of CD34+ HPC. In spite of these differences a significant degree of correlation (r > 0.83 in all cases) was found between both methods regarding the total number of CD34+, CD34+/CD19- and CD34+/CD19+ HPC present in the BM, PB and PB-derived leucapheresis samples analyzed in this study.