Interference of blast cell fragments with automated platelet counting

Micro-Red Blood Cell, Fragmented Red Blood Cell, Platelet Distribution Width, Mean Platelet Volume, and Platelet-Large Cell Ratio on Sysmex XN Series Hematology Analyzers Can Be Used for the Reflex Test of Impedance Platelet Count in Clinical Practice

CONTEXT.—

Platelet (PLT) counting with impedance (PLT-I) is widely used but has low specificity. PLT counting with fluorescence (PLT-F), tested by the Sysmex XN series with high specificity, can be a complementary method to PLT-I.

OBJECTIVE.—

To identify red blood cell (RBC)- and PLT-related parameters as potential influencing factors for PLT-I and establish PLT reflex test rules with PLT-F.

DESIGN.—

We prospectively tested both PLT-I and PLT-F in all 3480 samples. In a development data set of 3000 samples, differences between the reflex and nonreflex groups were compared and influencing factors for PLT-I were identified by logistic regression. The area under the receiver operating characteristic (ROC) curve and cutoff values were obtained by ROC curve analysis. Validation was conducted in the remaining 480 samples (validation data set).

RESULTS.—

PLT-F showed comparable results with immunoplatelet counting. In logistic regression, increased micro-RBC absolute count (micro-RBC#), fragmented RBC absolute count (FRC#), PLT distribution width (PDW), mean PLT volume (MPV), PLT-large cell ratio (P-LCR), and immature PLT fraction absolute count (IPF#) were influencing factors for PLT-I. In ROC curve analysis, the cutoff values of micro-RBC#, FRC#, PDW, MPV, and P-LCR were 0.64 × 106/μL, 0.082 × 106/μL, 15.40 fL, 11.15 fL, and 33.95%, respectively. The areas under the ROC curve of micro-RBC# and FRC# were 0.77 and 0.79, respectively.

CONCLUSIONS.—

Micro-RBC#, FRC#, PDW, MPV, P-LCR, and IPF# were factors affecting PLT-I. Among them, micro-RBC# and FRC# were the most impactful factors. From our study results, micro-RBC#, FRC#, MPV, PDW, and P-LCR can be used to establish reflex test rules for PLT counting in clinical work.

Platelet count in impedance-based hematology analyzer: Beware of trap!

Among the blood counts, platelet count is most often reported with inconsistency. Many of the analyzers work on electrical impedance principle for red blood cell (RBC) and platelet counting. However, with this technology, factors such as fragmented RBCs, microcytes, cytoplasmic fragments of leukemic cells, lipid particles, fungal yeast forms, and bacteria are known to interfere with platelet count and give spuriously elevated platelet counts. A 72-year-old male was admitted for the treatment of dengue infection who had serial platelet count monitoring. He had an initial platelet count of 48,000/cumm which suddenly improved to 2.6 lakhs within 6 h without any platelet transfusion. Peripheral smear however did not correlate with the machine-derived count. Repeat test after 6 h yielded a result of 56,000/cumm which correlated well with the peripheral smear. This falsely elevated count was due to the presence of lipid particles as the sample was drawn in the postprandial state.

Analytical performance of automated platelet counts and impact on platelet transfusion guidance in patients with acute leukemia

The aim of this study was to evaluate the performance of automated impedance platelet counts by Beckman Coulter LH780 (PLT-LH), Sysmex XN-3000 (PLT-XNi) and fluorescence method by Sysmex XN-3000 (PLT-F) in patients with acute leukemia. Blood specimens were subjected to platelet measurements by evaluated methods and then compared against the international reference method (IRM). Eighty-two blood specimens were included. Bland-Altman plots of the differences between the evaluated methods and IRM demonstrated mean biases of PLT-LH, PLT-XNi and PLT-F of 9 × 10⁹/L, 11 × 10⁹/L and 2 × 10⁹/L, respectively. For platelet transfusion guidance, all evaluated methods had acceptable accuracy. For platelet transfusion guidance, the sensitivities of PLT-LH, PLT-XNi and PLT-F were 33.3, 25.0 and 83.3%, respectively, at a transfusion threshold of 10 × 10⁹/L, and 73.1, 61.5 and 84.6%, respectively, at transfusion threshold of 20 × 10⁹/L. High blast count was associated with inaccurate PLT-LH and PLT-XNi. In conclusion, the PLT-F demonstrated excellent performance for diagnosis of thrombocytopenia and for platelet transfusion guidance in the evaluated specimens from acute leukemia patients. With respect to clinical relevance, careful blood smear review is necessary in case of high blast counts.