Crossover study of three methods for low-level white blood cell counting on two types of flow cytometer

Background: Flow cytometric methods were previously shown to be preferable to microscopic and volumetric methods for counting residual white blood cells (WBCs). In this study, three flow cytometric, low-level WBC counting methods were cross compared using two flow cytometers. Methods: Double-filtered red cell and platelet concentrates were spiked with different amounts of WBC to obtain panels of unspiked and 0.3, 1.0, 3.3, and 10.0 WBC/mul. The methods of BD Biosciences (BOB), Beckman-Coulter (BC), and an in-house method were performed on flow cytometers from BOB and BC. Samples were measured in ninefold. We required that (a) r(2) be at least 0.98 (linearity), (b) at least 80% of observations fell within 20% of expected values (accuracy), and (c) the coefficients of variation be at least 20% (precision) for samples containing at least 3.3 WBC/mul. Results: For the red cell panel, our requirements were met by the BOB method on both flow cytometers and by the BC and in-house methods on the BOB flow cytometer only. For the platelet panel, our requirements were met on all combinations of methods and flow cytometers, except for the in-house method on the BOB flow cytometer. Intra-assay variation was lowest for the BOB method, irrespective of the type of flow cytometer used. Conclusion: Based on accuracy and precision, the BOB method on the BOB flow cytometer produced the best results for counting low-level WBCs. (C) 2003 Wiley-Liss, Inc.