Manual and automated methods for the determination of leukocyte counts at extreme low levels: comparative evaluation of the Nageotte chamber and the Abbott Cell Dyn 3500 analyser.

Leukodepleted or leukocyte-poor blood products (fresh-frozen plasma, packed red cell and platelet concentrates in particular) are widely used in current clinical practice. However, because the monitoring of leukodepletion efficiency is generally carried out (if at all) using the labour-intensive and relatively inaccurate manual Nageotte chamber technique, it is clear that any increased demand for leukodepletion monitoring would be difficult, if not impossible, to meet. As the need to identify an automated alternative to the Nageotte technique is important, this study was undertaken to evaluate such a possibility. White blood cells were enumerated in a representative series of filtered and non-filtered human blood components by both microscopic counting in the Nageotte chamber, and with the Abbott CD3500 automated haematology analyser. For the Nageotte estimate, a single analysis was made in accordance with standard procedures, whereas the automated analysis was achieved by making six replicate counts and determining the mean of four replicates after excluding the highest and lowest estimates. To determine linearity limits of the manual and automated procedures, freshly isolated leukocytes were admixed with cell-free plasma-pheresis plasma. Reasonable reproducibility (mean CV 10% for cell counts exceeding 100 cells/microL) and good linearity (r > 0.9) were observed for CD3500 determinations in four separate experiments. The manual and automated measurements also correlated well (r > 0.9) with no obvious inter-method bias for cell counts up to 40 cells/microL although there was some suggestion of lower absolute CD3500 counts in the range 40-130 cells/microL. For the comparative studies with filtered and non-filtered blood products, no significant method bias was seen with 70 individual red cell concentrates, but systematically higher CD3500 white blood cell counts were observed in the series of 68 platelet concentrates (probably due to the presence of platelet clumps). This study concludes that automation of white cell counts in blood products with the CD3500 analyser is feasible for quality control in the preparation of fresh-frozen plasma and red cell concentrates but is limited for the analysis of filtered platelet concentrates.