Automated Platelet Counts and Transfusion in Cancer Patients
Automated Platelet Counts and Transfusion in Cancer Patients
This study was designed to evaluate the accuracy and precision of contemporary automated platelet-counting methods in patients with cancer who have associated thrombocytopenia. Our findings of inaccuracies when using optical and impedance platelet counts are consistent with prior reports. In addition, the percentage of severely thrombocytopenic patients who may not have received a needed transfusion because of falsely elevated platelet counts was higher than in most previous reports.
The potential effect on transfusion decisions demonstrated in this study is, to our knowledge, one of the largest reported to date. More than 40% of patients who are transfusion eligible may indeed not be receiving one if strict guidelines are followed and optical/impedance methods are used. The risk of undertransfusion greatly increased as the threshold was lowered below 10 × 10/L, and the error rate differed depending on the analyzer/method used in the platelet determination. Based on these data, lowering the platelet threshold to 5 × 10/L or less may be problematic given the technical issues reported above.
In an attempt to further examine the potential cause(s) for the falsely increased platelet counts seen in this study, peripheral blood smears were retrieved in 10 discrepant cases in which the CD61 counts were less than 10 × 10/L. In most of the smears examined, there were no appreciable numbers of RBC fragments, microcytes, WBC cytoplasmic fragments, or other nonplatelet particles that may have falsely elevated the automated platelet counts.
We believe that this study is important because it introduces the accuracy of measuring low platelet number as a variable for consideration at a time when large studies are appearing in the literature that advocate using either no or lower prophylactic transfusion triggers in distinct subpopulations of patients with severe thrombocytopenia. This study is extensive in that it presents one of the largest data sets to address this question in a prospective manner. In addition, there is clinical correlation for a subset of patients as to the timing of the platelet transfusion as a response to the laboratory-based trigger. The study was performed at a major tertiary care center where thrombocytopenia is relatively common. Many of the patients at MSKCC are transfusion eligible, and while a trigger of 10 × 10/L or less has been recommended, decision algorithms used by clinicians can vary, and the data appear to support this. Furthermore, it is unlikely that these decisions are made in light of understanding the potential for inaccuracy in the platelet count. The biases we found have potential clinical implications since the standards for platelet transfusion are undergoing revision. Because of this, we believe it is important to introduce this as a potential confounder and help provide a critical assessment of how the laboratory triggers, which form the basis of clinical care, are generated.
Clinical laboratories (and those who design and develop instrumentation) strive to produce automated platelet counts that are both accurate and precise throughout the reportable range of the analyzer. This presents a problem in severely thrombocytopenic samples since it has been demonstrated that clinically significant biases may affect decisions regarding when to administer platelet transfusions. What, then, are the alternatives? Laboratories could shift to an alternate method when faced with these problematic samples. The manual phase platelet count, while recognized as the gold standard for the past 50 years, is not a feasible alternative because it is time-consuming, resource draining, and technically difficult, and it suffers from very poor precision. Should laboratories report a platelet count obtained through smear estimation? While this may be a quick and easy alternative, it is neither accurate nor precise. Such problems generally prohibit its use except to estimate a platelet count as decreased, normal, or increased. Immunologic platelet-counting methods based on flow cytometric quantitation of platelets labeled with anti-CD41 and CD61 have proven to be extremely accurate and reproducible, even in severely thrombocytopenic samples. It has been suggested that the immunologic platelet count is a suitable alternative to optical and impedance methods in severely thrombocytopenic samples, especially when deciding whether a platelet transfusion should be administered. Kunz et al have suggested that platelet counts close to transfusion trigger thresholds should be verified with an immunologic method.
Widespread adoption of immunologic methods of platelet determinations poses several important challenges. In many hospitals, the flow cytometry and hematology laboratories are physically separated from each other and operate independently. The flow cytometry laboratory would need to develop and implement CD41/61 immunoplatelet counting and incorporate it into its workflow. In a large cancer center setting, this added volume would be difficult to accommodate since the priority of the flow cytometry laboratory is the diagnosis and monitoring of hematologic neoplasms. Further complicating such a practice is that the turnaround times and logistics associated with transporting, processing, and reporting the flow cytometric platelet count may prove to be unacceptably long for clinicians needing to make quick decisions regarding patient assessment and treatment. Alternatively, an automated immunologic platelet count is currently available on one of the hematology analyzers tested in this study; however, the test cost is similar to immunophenotyping assays, processed as a separate test and not routinely done as part of the instrument's complete blood cell count.
Therefore, the laboratory may choose to be selective in determining which samples require specialized testing. With the immunoplatelet method, one can envision a "two-tiered" approach in which samples are initially screened with optical/impedance analyzers, and those patients who have been determined to have severe thrombocytopenia have the precise level verified by immunologic techniques. Further prospective studies may be required to determine whether such testing is clinically indicated for most patients with malignant disease.
Discussion
This study was designed to evaluate the accuracy and precision of contemporary automated platelet-counting methods in patients with cancer who have associated thrombocytopenia. Our findings of inaccuracies when using optical and impedance platelet counts are consistent with prior reports. In addition, the percentage of severely thrombocytopenic patients who may not have received a needed transfusion because of falsely elevated platelet counts was higher than in most previous reports.
The potential effect on transfusion decisions demonstrated in this study is, to our knowledge, one of the largest reported to date. More than 40% of patients who are transfusion eligible may indeed not be receiving one if strict guidelines are followed and optical/impedance methods are used. The risk of undertransfusion greatly increased as the threshold was lowered below 10 × 10/L, and the error rate differed depending on the analyzer/method used in the platelet determination. Based on these data, lowering the platelet threshold to 5 × 10/L or less may be problematic given the technical issues reported above.
In an attempt to further examine the potential cause(s) for the falsely increased platelet counts seen in this study, peripheral blood smears were retrieved in 10 discrepant cases in which the CD61 counts were less than 10 × 10/L. In most of the smears examined, there were no appreciable numbers of RBC fragments, microcytes, WBC cytoplasmic fragments, or other nonplatelet particles that may have falsely elevated the automated platelet counts.
We believe that this study is important because it introduces the accuracy of measuring low platelet number as a variable for consideration at a time when large studies are appearing in the literature that advocate using either no or lower prophylactic transfusion triggers in distinct subpopulations of patients with severe thrombocytopenia. This study is extensive in that it presents one of the largest data sets to address this question in a prospective manner. In addition, there is clinical correlation for a subset of patients as to the timing of the platelet transfusion as a response to the laboratory-based trigger. The study was performed at a major tertiary care center where thrombocytopenia is relatively common. Many of the patients at MSKCC are transfusion eligible, and while a trigger of 10 × 10/L or less has been recommended, decision algorithms used by clinicians can vary, and the data appear to support this. Furthermore, it is unlikely that these decisions are made in light of understanding the potential for inaccuracy in the platelet count. The biases we found have potential clinical implications since the standards for platelet transfusion are undergoing revision. Because of this, we believe it is important to introduce this as a potential confounder and help provide a critical assessment of how the laboratory triggers, which form the basis of clinical care, are generated.
Clinical laboratories (and those who design and develop instrumentation) strive to produce automated platelet counts that are both accurate and precise throughout the reportable range of the analyzer. This presents a problem in severely thrombocytopenic samples since it has been demonstrated that clinically significant biases may affect decisions regarding when to administer platelet transfusions. What, then, are the alternatives? Laboratories could shift to an alternate method when faced with these problematic samples. The manual phase platelet count, while recognized as the gold standard for the past 50 years, is not a feasible alternative because it is time-consuming, resource draining, and technically difficult, and it suffers from very poor precision. Should laboratories report a platelet count obtained through smear estimation? While this may be a quick and easy alternative, it is neither accurate nor precise. Such problems generally prohibit its use except to estimate a platelet count as decreased, normal, or increased. Immunologic platelet-counting methods based on flow cytometric quantitation of platelets labeled with anti-CD41 and CD61 have proven to be extremely accurate and reproducible, even in severely thrombocytopenic samples. It has been suggested that the immunologic platelet count is a suitable alternative to optical and impedance methods in severely thrombocytopenic samples, especially when deciding whether a platelet transfusion should be administered. Kunz et al have suggested that platelet counts close to transfusion trigger thresholds should be verified with an immunologic method.
Widespread adoption of immunologic methods of platelet determinations poses several important challenges. In many hospitals, the flow cytometry and hematology laboratories are physically separated from each other and operate independently. The flow cytometry laboratory would need to develop and implement CD41/61 immunoplatelet counting and incorporate it into its workflow. In a large cancer center setting, this added volume would be difficult to accommodate since the priority of the flow cytometry laboratory is the diagnosis and monitoring of hematologic neoplasms. Further complicating such a practice is that the turnaround times and logistics associated with transporting, processing, and reporting the flow cytometric platelet count may prove to be unacceptably long for clinicians needing to make quick decisions regarding patient assessment and treatment. Alternatively, an automated immunologic platelet count is currently available on one of the hematology analyzers tested in this study; however, the test cost is similar to immunophenotyping assays, processed as a separate test and not routinely done as part of the instrument's complete blood cell count.
Therefore, the laboratory may choose to be selective in determining which samples require specialized testing. With the immunoplatelet method, one can envision a "two-tiered" approach in which samples are initially screened with optical/impedance analyzers, and those patients who have been determined to have severe thrombocytopenia have the precise level verified by immunologic techniques. Further prospective studies may be required to determine whether such testing is clinically indicated for most patients with malignant disease.
