Comparison of the effects of Ficoll-Hypaque separation and whole blood lysis on results of immunophenotypic analysis of blood and bone marrow samples from patients with hematologic malignancies.

We compared flow cytometric immunophenotyping results obtained by using the lysed whole blood method of sample preparation with those obtained by using Ficoll-Hypaque-separated cells on 44 consecutive specimens from patients with various hematologic malignancies. When the samples were analyzed as a group, seven antigens (CD2, CD3, CD5, CD11c, CD20, CD22, and CD34) demonstrated significantly different percentages of positively staining cells. When the samples were grouped by disease, results for patients with acute lymphocytic leukemia were discordant for CD22 and HLA-DR and results for patients with hairy cell leukemia were discordant for CD34. Most of the differences, however, were not with antigens critical to the evaluation of the malignancy. Additionally, the most frequent reason for differences in the percentage of positive cells was due to isotype control-based placement of the quadrant markers and not an actual discrepancy in staining. However, analysis of the CD34 antigen yielded eight instances in which staining of Ficoll-Hypaque-separated cells was essentially negative, but a clearly positive population was evident with the lysed preparation. This finding has important implications because of the prognostic significance of this antigen. Further studies are needed to determine the cause of this phenomenon.