Standardisation of platelet counting accuracy in blood banks by reference to an automated immunoplatelet procedure: comparative evaluation of Cell-Dyn CD4000 impedance and optical platelet counts.

Prophylactic and therapeutic platelet transfusions are increasingly used for patients with conditions associated with thrombocytopenia in order to prevent the development of potentially life threatening bleeding. These clinical strategies have led to a significant expansion in platelet unit manufacture, and this now represents a major resource and cost commitment for blood banks. As part of the manufacturing process, blood banks are required to implement control procedures, and the determination of platelet counts in particular is necessary to confirm that the quality of platelet unit production meets the standards defined by national or international guidelines. Apart from linearity analysis and comparisons of platelet counts given by different instruments, there has been no systematic standardisation of platelet counting methods in blood bank practice because to date there has been no suitable reference method for counting platelets in citrate anticoagulants. The recent introduction of an automated immunoplatelet procedure on the Cell-Dyn CD4000 provides a means of determining a true platelet count that is unaffected by changes induced either by storage or anticoagulant. The CD4000 in its routine configuration also provides simultaneous impedance and optical platelet counts and this study was therefore undertaken in order to compare all three different platelet counting methods in parallel with a representative series of platelet units. Platelet counts determined after sub-sampling of platelet units into EDTA vs plain non-anticoagulated tubes revealed no differences in impedance or immunoplatelet counts but generally lower optical counts when aliquoted into tubes that did not contain EDTA. This study therefore routinely used EDTA for platelet unit sub-samples. Comparative results of platelet counts for buffy coat platelet units (n = 36) aliquoted into EDTA indicated that the impedance count was higher than the reference immunoplatelet count by a mean factor of 1.25 while the optical count was lower by a mean factor of 0.87. The degree of impedance count overestimation was particularly consistent while the optical count underestimation was more variable. Linearity studies of 10 fresh platelet units showed no deviation in the range 0-2305 x 10(9) l(-1) for impedance and 0 to 1420 x 10(9) l(-1) for the optical counts, and the relative numerical relationships between impedance and optical counts were conserved throughout the range of dilutions tested. In the CD4000 optical analysis, blood samples anticoagulated with EDTA showed a distinctive elliptical population distribution that fell within the system thresholds. In contrast, the optical pattern observed for platelet units (in CPD) and ACD-anticoagulated venous blood showed a wider 90 degrees scatter with a population of platelet events above the upper parallel discriminator. As these were excluded from the optical count (but were still identified as platelets by the immunoplatelet method) it meant that the optical counts of samples in citrate-based anticoagulants were systematically lower than immunoplatelet counts. Platelet units (n = 15) analysed daily over a seven day period of storage revealed that the greatest decline in platelet counts was with the optical measurement while the most stable value was obtained by impedance analysis. The results of the immunoplatelet analysis further suggested a progressive increase in small platelets with increasing storage time. The use in this study of a standardised immunoplatelet reference method to examine the question of analyser suitability for determining platelet counts/yields of platelet units thus provided a number of important findings. An impedance platelet counting method is utilised by the great majority of haematology instruments in current use, and in common with the CD4000 analyser, a correction factor is employed to take account of RBC/platelet coincidence. This study found that when analysed samples such as platelet units were RBC-free, that an inappropriate correction factor was applied. Consequently, the CD4000 impedance platelet count will provide reliable platelet counts, irrespective of the day of platelet unit storage, when a factor of 1.25 is applied to the system-reported result. By comparison, optical methods are more likely to be affected by subtle morphological changes that may result from anticoagulants or progressive storage time. The method limitations documented by this study may well affect many other analysers and mean that the implementation of process control statistics related to platelet counts may be less reliable than previously assumed. It is suggested that standardisation could be much better achieved if there was some form of system cross-calibration that was referenced to an independent method such as an immunoplatelet assay. It is proposed that studies of this type should be extended to a wide assessment of platelet count accuracy of blood bank instruments in order to standardise data within national organisations. If consistent inter-instrument correction factors such as those documented here can be identified, it would considerably increase the relevance of determining platelet counts in production control processes.