Potential Clinical Impact of Inaccurate Automated Platelet Counts in the Setting of Severe Thrombocytopenia

Abstract 3428

Severe thrombocytopenia is not an uncommon finding in the setting of hematologic malignancies. The key to a successful outcome often lies in the ability to support the patient through periods of cytopenia by adopting an aggressive transfusion strategy for platelets. Therefore, an accurate assessment of the platelet count is of utmost importance particularly as it relates to various levels which “trigger” prophylactic platelet transfusion. However, the accuracy and precision of conventional platelet counting methods have been shown to be suspect in severely thrombocytopenic samples. We investigated the accuracy of several commercial hematology analyzers currently available in clinical practice which use optical, impedance and immunologic platelet determination methods and assessed whether biased results could have an impact on decisions regarding platelet transfusion. Four-hundred three (403) EDTA-anticoagulated samples collected from patients with hematologic malignancies and a platelet count of < 50 × 10e3/uL were selected from the standard workflow processed in the hematology laboratories at MSKCC. Samples were split and tested within 8 hours of collection using optical based methods on the Advia 2120i (Siemens, Tarrytown, NY) and CELL-DYN Sapphire (Abbott Diagnostics, Santa Clara, CA). During the study, an additional analyzer became available, the XE-2100 (Sysmex, Kobe, Japan) which enumerates platelets by impedance and optical technology. One hundred twenty seven (127) of the 403 samples were split and tested by the Sysmex analyzer. Platelet counts from each method were then compared using the CD61 immunoplatelet determination (Abbott, Santa Clara, CA) as the reference value. The CD61 immunoplatelet method was chosen as the reference standard because we found excellent correlation in linear regression between this method and phase microscopy (r =0.99, y = −0.96 + 0.88x; n = 37). Also, the International Counsel for Standardization in Haematology and the International Society for Laboratory Hematology recommend the counting of specifically labelled platelets (CD41, CD61) by flow cytometry as a reference method for the enumeration of platelets. Bland-Altman difference plots were used to visualize the agreement between the reference and test methods and the paired t-test evaluated the statistical significance of the difference between methods. We then compared the number of platelet transfusion indications (at various platelet thresholds) as determined by all methods with potential transfusion decisions made using the reference method. There was a statistically significant positive bias among the optical and impedance methods compared with CD61 enumeration. Using various platelet transfusion decision points, the number of patients who were at risk for under transfusion (platelet count above threshold when reference result is < threshold) ranged from 1–2 % at a threshold of 50k/uL to 30–40 % when the threshold was lowered to 10k/uL. These results highlight the limitations in the accuracy of optical and impedance methods, particularly in the setting of severe thrombocytopenia. Since prophylactic platelet transfusions are by and large based on standard “triggers,” an overestimation of the platelet count may lead to under-transfusion, especially in patients at the highest risk for hemorrhage. Clinicians should be aware of the limitations of automated platelet determinations and consider this when making medical decisions regarding supportive transfusions.