Chaicharoen Tantanate

Performance evaluation of the automated nucleated red blood cell enumeration on Sysmex XN analyser

Presence of peripheral blood nucleated red blood cell (NRBC) is associated with pathological conditions and leads to the overestimation of white blood cell count in automated haematology analysers (HA). The authors evaluated NRBC enumeration by a new HA Sysmex XN (XN) to demonstrate the precision and comparability to manual count (MC) at the various NRBC values.

Spurious white blood cell count from a new automated Sysmex XN hematology analyzer.

Sir, Until recently, most clinical laboratories have implemented automated hematology analyzers (HA) for complete blood count (CBC). Although new technologies have been developed, spurious cell counts can occur [1, 2]. The white blood cell (WBC) count is one of the important CBC parameters. Spurious WBC count may lead to inappropriate treatment and unnecessary investigation. Causes of spurious WBC count may relate to patient’s condition, type, and amount of used anticoagulant or abnormality of analytical process of HA. The author accidentally encountered a case with an incorrect report of WBC count from Sysmex XN (XN) analyzers (Sysmex Corporation, Kobe, Japan). A 56year-old female came to Siriraj Hospital in October 2013 for the follow-up of hypertension and dyslipidemia. Her medications included 50 mg/day of losartan, 5 mg/day of amlodipine, and 10 mg/day of simvastatin. Her blood pressure was well under controlled. She had no abnormal symptom and her physical examinations were within normal limits. CBC and blood chemistries were requested for her annual checkup. Her previous CBC about a year ago, analyzed by Sysmex XE-5000 (Sysmex Corporation), showed normal results. For this latest visit, her CBC was performed by the new XN. The results revealed hemoglobin 110 g/L, mean corpuscular volume 83.2 fL, platelet count 313 9 10/L, WBC count 0.11 9 10/L, neutrophil 0.06 9 10/L, lymphocyte 0.04 9 10/L, and monocyte 0.01 9 10/L. The alert flags of leucopenia, neutropenia, and lymphopenia were sent from the XN to the laboratory information system (LIS), and a blood smear was then automatically prepared and reviewed according to the institutional policy. It was found that the WBC in blood smear appeared higher than that reported by the XN without any aggregation. Nucleated red blood cells (NRBCs) were not detected. The specimen was analyzed again by another XN, and the result was similar with a WBC count of 0.13 9 10/L. To confirm the WBC count, the Beckman Coulter LH780 (Beckman Coulter, Miami, FL, USA) was used to reanalyze the sample. The WBC count from this HA was 7.4 9 10/L without any flag. This confirmed the estimated WBC in the smear, and therefore, the actual WBC count was not as low as reported by the XN. The author searched for the cause of this occurrence. The scattergrams were reviewed and it was found that the total nucleated cell count (TNC) in the WNR channel was very low (0.11 9 10/L) while that in the WDF channel was quite normal (Figure 1). The TNC in WDF channel was 6.93 9 10/L. This count again was more compatible with the findings in blood

Current practices of bleeding time in a developing country: an alert for noncompliance with the standard procedures.

Bleeding time is still widely performed in many developing countries including Thailand. To generate an accurate result, the procedure should be complied with standard recommendations such as those from Clinical and Laboratory Standards Institute (CLSI) and World Federation of Hemophilia (WFH). The authors surveyed the current practices of bleeding time in Thailand in order to verify the practices that did not comply with the accepted standard.

International collaborative study for the calibration of proposed International Standards for thromboplastin, rabbit, plain, and for thromboplastin, recombinant, human, plain

Essentials Two candidate International Standards for thromboplastin (coded RBT/16 and rTF/16) are proposed. International Sensitivity Index (ISI) of proposed standards was assessed in a 20‐centre study. The mean ISI for RBT/16 was 1.21 with a between‐centre coefficient of variation of 4.6%. The mean ISI for rTF/16 was 1.11 with a between‐centre coefficient of variation of 5.7%.

Spuriously high platelet counts by various automated hematology analyzers in a patient with disseminated intravascular coagulation

Until recently, the platelet enumeration by automated hematology analyzers (HAs) has been widely accepted for routine testing. This method replaces the traditional manual counting in many laboratories because of its better precision and convenience. Various principles of HAs have been applied for platelet count, including electrical impedance, light scattering, fluorescent detection, and immunoplatelet count [1]. Although the technologies have been continuously improved, the spurious platelet counting can occur and may lead to an inappropriate patient management. The spurious count can be either higher or lower than the actual value depending on the conditions of the patients and the principles of HAs. The conditions related with this phenomenon have been reviewed extensively by Zandecki et al. [2]. The author examined a blood sample with a spuriously high platelet count. This sample came from a 1-year-old girl with underlying diseases of biliary atresia and cirrhosis. She was admitted with high grade fever and diarrhea. The diagnosis was infectious diarrhea with Klebsiella pneumonia septicemia. During the admission, she developed abnormal bleeding including petechiae, ecchymosis, and bleedings at the puncture and catheter sites. Her CBC performed by a routine mode on the Sysmex XN-3000 (XN-3000; Sysmex Corporation, Kobe, Japan) revealed hemoglobin 92 g/L, mean corpuscular volume 69 fL, WBC count 14.7 × 109/L, neutrophil 8 × 109/L, lymphocyte 2.7 × 109/L, monocyte 1.9 × 109/L, and platelet count 193 × 109/L. The blood smear is shown in Figure 1. The important findings were red blood cell (RBC) abnormalities including numerous schistocytes, microspherocytes, and some polychromasias. It is noted that the platelet estimation from the blood smear was decreased (1–2 platelets per oil field) and this was discordant with that from the HA. The screening coagulogram revealed APTT 107.6 s, PT 31.6 s, and fibrinogen 1.53 g/L (the reference ranges of APTT, PT, and fibrinogen were 23–32 s, 11–13 s, and 1.5–4 g/L, respectively). The D-dimer was 7122 μg/L (reference range: 68–494 μg/L). According to the International Society on Thrombosis and Haemostasis (ISTH) scoring system for overt disseminated intravascular coagulation (DIC), the score was at least 5 due to the prolonged PT (2 points) and markedly increased D-dimer (3 points) [3]. Therefore, the diagnosis was an acute overt DIC. Although the platelet value did not change the diagnosis in this case, the accurate count was still crucial for the treatment guidance and follow-up. The author further investigated the CBC with other methods, including electrical

Evaluation of Thrombi-Stat MC1 for whole blood and plasma international normalized ratio in comparison with a laboratory method

The precision and comparability of the international normalized ratio (INR) analyzed by the Thrombi‐Stat MC1 coagulation testing system were evaluated before using as a point‐of‐care testing.

Performance Evaluation of High Fluorescence Lymphocyte Count: Comparability to Atypical Lymphocyte Count and Clinical Significance

Objectives To investigate the association between high-fluorescence lymphocyte cell (HFLC) and atypical lymphocyte (AL) counts, and to determine the clinical significance of HFLC. Methods We compared automated HFLC and microscopic AL counts and analyzed the findings. Patient clinical data for each specimen were reviewed. Results A total of 320 blood specimens were included. The correlation between HFLC and microscopic AL counts was 0.865 and 0.893 for absolute and percentage counts, respectively. Sensitivity, specificity, and accuracy of HFLC at the cutoff value of 0.1 × 109 per L for detection of AL were 0.8, 0.77, and 0.8, respectively. Studied patients were classified into 4 groups: infection, immunological disorders, malignant neoplasms, and others. Patients with infections had the highest HFLC. Most of those patients (67.7%) had dengue infection. Conclusion HFLC counts were well-correlated with AL counts with the acceptable test characteristics. Applying HFLC flagging may alert laboratory staff to be aware of ALs.

Performance Evaluation of Automated Impedance and Optical Fluorescence Platelet Counts Compared With International Reference Method in Patients With Thalassemia

CONTEXT - Spurious platelet counts from automated methods have been reported in patients with abnormal red blood cells. However, there is no specific study regarding performance of platelet counts by automated methods in patients with thalassemia. OBJECTIVE - To investigate the performance of automated platelet counts, including impedance (PLT-I) and optical fluorescent (PLT-O and PLT-F) methods, and compare them with the international reference method (IRM) for platelet counting in patients with thalassemia. DESIGN - Two hundred forty-nine thalassemia specimens from various subtypes were examined. PLT-I, PLT-O, and PLT-F from a Sysmex XN analyzer were evaluated and compared against the IRM. Demographic data, platelet counts, and red blood cell parameters are shown. Comparability between evaluated methods and IRM, as well as test characteristics, is presented. Factors involving inaccurate PLT-I were analyzed. RESULTS - Specimens with platelet counts ranging from 31 × 103/μL to 932 × 103/μL were included. Most patients were patients with thalassemia major. Correlation between PLT-I and IRM was lower than that of the other methods in overall patients. PLT-O and PLT-F were correlated to IRM when classifying patients according to clinically significant platelet ranges. All automated methods had acceptable sensitivities; however, specificity of PLT-I was low for diagnosis of thrombocytopenia. High RDW-CV (red blood cell distribution width-coefficient of variation) was an independent factor of inaccurate PLT-I measurement. CONCLUSIONS - Among the evaluated methods, PLT-I was the method least correlated to IRM, with PLT-O and PLT-F comparable to IRM in patients with thalassemia. Optical platelet counts and careful blood smear examination are recommended alternative platelet counting methods, depending on the clinical setting.

Response to the comments on the article ‘Performance evaluation of automated nucleated red blood cell enumeration on Sysmex XN analyser’ (IJLH-10-14-0363)

Sir, The author would like to thank for the comments from Dr. Hoffmann and Dr. van Dun on the article ‘Performance evaluation of automated NRBC enumeration on Sysmex XN analyser’ [1]. The author would like to respond as these followings. About the comment about sample selection, the specimens with more than 0.1% NRBCs by the Sysmex XN3000 (XN) were selected in this study because the author had to collect most of the specimens of interest but the manual microscopic count (MC) was not performed in every specimen. However, the cut-off value of positive result at the level of NRBCs more than 1% was used to analyse the false negative (FN) as mentioned in the article. This cut-off was used because some institutes apply this value as a threshold for reporting NRBC and correcting the numbers of white blood cells. The author did not use the level of 0.1% for the analysis, so the selection bias could be reduced. The author made a conclusion about the FN and false-positive (FP) reports by the XN based on the data in the Table 1. There was no FN by the XN because all specimens with NRBCs <1% by the XN also had <1% NRBC by the MC. The FP was also not observed, as demonstrated in the Table 1. However, the author agrees that there were some specimens with more NRBCs by the XN compared to the MC, as demonstrated in the Figure 3, especially when looking at the specimen with NRBCs more than 1%. Further study to analyse the sensitivity and specificity of XN compared to MC could be done to obtain more information. The author performed further analysis about the correlation between XN and MC. Because the data were not normally distributed, the Spearman’s Rank Correlation was selected. The rs was 0.99 which demonstrated an excellent correlation [2]. The intraclass correlation coefficients (ICC) were also analysed; the ICC between XN and MC were 0.98 (95% CI, 0.98–0.99) which confirmed the excellent reliability between methods [3]. When calculating the Passing–Bablok regression analysis according to the commentators’ recommendation, there was no identity between methods as demonstrated by the equation y = 0.30 + 0.80x (95% CI of slope were 0.77–0.82). This finding along with the Bland-Altman plot in Figure 3 demonstrates that the NRBC counts by the XN were certainly not identical to those by the MC for all ranges of NRBCs. The author also agrees with the comment about the bias between XN and MC. However, the author concluded ‘that XN could replace the traditional MC for NRBC, MC, especially for the specimens with NRBCs lower than 200%’ based on the concordance analysis of the data in the Table 1. This kind of analysis has also been used in the study by Kwon et al. [4]. The concordance rate between XN and MC at the NRBCs <200% was high. However, one of the limitations of the study was the high imprecision of the MC as demonstrated in Figure 1. Therefore, the author agrees with the commentators that further studies with solid design about patient selection and using more sensitive and specific assay as a comparative method should be performed. Importantly, the studies should include a wide range of NRBC numbers in order to be able to apply the knowledge to the laboratories in an area with high prevalence of chronic haemolytic anaemias.