|
|
| Year : 2013 | Volume
: 56
| Issue : 2 | Page : 114-119 |
|
| Comparison of platelet counts by sysmex XE 2100 and LH-750 with the international flow reference method in thrombocytopenic patients |
|
|
Tina Dadu, Kunal Sehgal, Anjum Shaikh, Shanaz Khodaiji
Department of Hematology, P. D. Hinduja Hospital & Medical Research Centre, Mumbai, Maharashtra, India
Click here for correspondence address and email
| Date of Web Publication | 23-Sep-2013 |
|
|
 |
|
 Abstract | | |
Background: There are several methods for counting platelets, of which the international flow reference method (IRM) is considered to be the gold standard. We compared the platelet count given by this method to the count given by automated analyzers using other methods, such as optical fluorescence and impedance. Aims: The aim of this study is to compare the platelet counts obtained by Sysmex XE 2100 by Impedance (Sysmex-I), optical florescence (Sysmex-O) and reported (Sysmex-R) based on the switching algorithm and LH-750 by Impedance (LH-750) with the IRM in thrombocytopenic blood samples. To calculate the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of various technologies at the clinically relevant transfusion thresholds of 10 × 10 9 /l and 20 × 10 9 /l. Materials and Methods: A total of 118 blood samples with platelet count of <50 × 10 9 /l were selected for the study. Platelet counts of all samples were analyzed by all methods using the Sysmex analyzer, LH-750 and IRM in parallel within 6 h of collection. Statistical Analysis Used: Pearson correlation, bland Altman analysis, sensitivity and specificity, PPV and NPV. Results and Conclusions: Sysmex-R had the least Bias and 95% limits of agreement (95%LA) range and thus correlated best with IRM values. LH-750 had a higher Bias compared to Sysmex-O and Sysmex-R, but a strikingly similar 95% LA ensures similar results in all three methods. In fact, in the oncology subset, it had the narrowest 95% LA, which made it the best performer in this subgroup. Of the three Sysmex results, Sysmex-I had the highest bias, widest 95% LA and highest potential risk of over transfusion. Hence, Sysmex-R and LH-750 were found to be reliable tools for estimation of platelet count in thrombocytopenic patients. Keywords: International flow reference method, LH-750, platelets, Sysmex XE 2100
How to cite this article:
Dadu T, Sehgal K, Shaikh A, Khodaiji S. Comparison of platelet counts by sysmex XE 2100 and LH-750 with the international flow reference method in thrombocytopenic patients. Indian J Pathol Microbiol 2013;56:114-9 |
How to cite this URL:
Dadu T, Sehgal K, Shaikh A, Khodaiji S. Comparison of platelet counts by sysmex XE 2100 and LH-750 with the international flow reference method in thrombocytopenic patients. Indian J Pathol Microbiol [serial online] 2013 [cited 2023 Sep 30];56:114-9. Available from: https://www.ijpmonline.org/text.asp?2013/56/2/114/118701 |
| Introduction | |  |
Precise platelet count is critical to a decision for transfusion. The decision to go in for platelet transfusion has to be weighed carefully. It could be a lifesaver where it is essential. The flip side though is that patients not requiring transfusion face a heightened risk from alloimmunization, immunosupression, infectious diseases and graft versus host disease. [1] This has led to efforts to reduce the threshold for platelet transfusion. Practice guidelines based on clinical outcome analysis suggest that the threshold for prophylactic platelet transfusion can be lowered from 20 10 9 /l to 10 10 9 /l or less for most clinical situations. [2],[3],[4],[5] However for this to be used resourcefully, a high degree of precision is required in platelet count below 20 10 9 /l.
International flow cytometry reference method (IRM) also known as the red blood cell (RBC) platelet ratio method is the current gold standard method provided by the international council for standardization in hematology (ICSH). [6] Platelet enumeration by this method is the most accurate technique currently available, but presents considerable challenges since it is time and labor intensive vis-à-vis platelet count by automated hematology analyzers and hence not suitable for routine platelet counting. Robust hematology analyzers, which can be used for platelet enumeration, promise to fill this void.
Most of the hematology analyzers use impedance method for counting platelets. However in thrombocytopenic patients this may or may not be the accurate count. New methodologies like optical method introduced by Sysmex and improved impedance method using "platelet fitting curve" introduced by Beckman Coulter try to make platelet counts more accurate. The aim of this study was to assess the accuracy and precision of different platelet counting technologies (impedance and optical florescence) by using different automated hematology analyzers, against the IRM in thrombocytopenic blood samples.
The analyzers evaluated are Beckman Coulter LH-750 and Sysmex XE 2100. The reference method used was the immunoplatelet RBC/Platelet ratio method.
| Materials and Methods | | |
A total of 118 blood samples with platelet count of <50 10 9 /l were selected for the study. Blood samples showing flags on either of the complete blood count analyzers (XE 2100 or LH-750) indicating the potential presence of platelet clumps or findings that potentially may interfere with automatic analysis were not included in this study. Samples were taken from the morning batch of samples received in the laboratory from admitted patients as well as from the out-patient department patients. Of the 118 blood samples selected, 45 were from oncology, 54 belonged to the infectious category and the rest 19 were put in the miscellaneous category (which included primarily idiopathic thrombocytopenic purpura (ITP), megaloblastic anemia, aplastic anemia, etc.).
Peripheral smears from all 118 samples were examined to check for the presence of platelet clumps. None of the slides showed any platelet clumping.
All blood specimens were collected in dipotassium ethylenediaminetetraacetic acid (K 2 EDTA) vacutainers. All samples were tested in parallel by different methods within 6 h of blood collection.
The instruments used were Sysmex XE 2100, which has the capability to provide both an impedance platelet count and a fluorescence-based optical platelet count when the sample is run in the reticulocyte mode and Coulter LH-750, which uses impedance technology with enhanced data extraction techniques and algorithms to eliminate Interference. For the IRM, FACS Canto II (Becton Dickinson, Franklin Lakes, NJ, USA) was used.
Platelet counts were thus evaluated in parallel by five methods:
- XE 2100-optical (Sysmex-O)
- XE 2100-impedance(Sysmex-I)
- XE 2100-reported (Sysmex-R) based on switching algorithm
- LH-750 impedance with curve fit correction on Beckman Coulter (LH-750)
- IRM done on the Becton Dickinson FACS Canto II (IRM).
Platelet count by IRM
Platelet count by IRM was based on the recommendation by the ICSH and the International Society of Laboratory Hematology. [6] The method is described in short here. 5 μl of the blood specimen is added to bovine serum albumin-phosphate-buffered saline (PBS-BSA) diluent, to which 5 μL of anti-CD41 and 5 μL of the anti-CD61 was added and incubated for 15 min. A final 1:1,000 dilution was made by adding 4.85 Ml of the PBS-BSA diluent. The antibody was mixed well with PBS-BSA by gentle in versions to ensure proper and equal distribution for counting of RBCs and platelets. With the flow cytometer, a minimum of 50,000 events with a minimum of 1,000 platelet events were counted. The final platelet count was derived using the following formula:
Platelet count = (*RBC count)/R
Where* RBC count represents the RBC cell count measured by LH-750 (impedance).
R = RBC gated events/platelet gated events.
Quality control
The instruments were calibrated according to the manufacturer's guidelines using the calibrators provided by them. Control samples (high, normal and low) were run as per College of American Pathologists (CAP) requirements.
Coefficient of variation (CV)
CV was calculated by running a single sample 11 times by each of the above five methods The mean platelet count of the sample used for evaluating the CV was 39 10 9 /l.
Statistical analysis
The data analysis was performed using the "Analyze IT" software. Pearson's correlation "r0" was used for evaluating the degree of correlation between different methods and the IRM. However, to compare two methods and find agreement between them Bland Altman (BA) method is the preferred method. [7] In Bland and Altman plot, the difference between the two measurements per subject is plotted against the mean of the two measurements. The limits of agreement (LA) are then defined as -1.96 s and +1.96 s (that describe the range for 95% of comparison points), with "s" the observed standard deviation of the difference between the two measurements per subject.
The positive predictive value (PPV) and negative predictive value (NPV) were also calculated to measure over transfusion and under transfusion rates.
| Results | | |
Quality control
The automated hematology analyzers and the IRM showed CV from 1.68% to 11.2% in precision studies.
The descriptive statistics and other details of the samples included in the study are shown in [Table 1].
A total of 118 K2EDTA-anti coagulated thrombocytopenic blood samples (<50 10 9 /l) were evaluated in this study. Of this, 22 had a count of <10 10 9 /l and 35 samples had account of <20 10 9 /l (including 22 samples, which had account of <10 10 9 /L).
Pearson correlation
The "r" value and slope equations for each method are shown in their representative graphs [Figure 1]. Platelet counts by all methods showed a very good correlation of >0.9 with IRM. | Figure 1: Correlati on of various methods with the internati onal flow reference method (platelet counts ×109/l)
Click here to view |
BA analysis
The mean differences and 95% LA were calculated for all the methods. These are plotted in [Figure 2]. | Figure 2: Bland-Altman plots showing the average bias and 95% limits of agreement of various methods with the international flow reference method (platelet counts ×109/l)
Click here to view |
The Bias was the highest for Sysmex-I (-2.805, 95% LA -15.857 to +10.246) followed by LH-750 (-2.636, 95% LA -12.966 to +7.695), Sysmex-O (+1.297, 95% LA -9.128 to +11.720) and least for Sysmex-R (+0.780, 95% LA -9.067 to +10.627). This indicates that on an average Sysmex-I substantially underestimates platelet counts compared with other methods and also had the widest 95% LA.
The widths of the 95% LA were also the widest for Sysmex-I. However, the widths of the 95% LA were equally wide for Sysmex-O, R and LH-750.
In 113 of 118 cases (95.7%), Sysmex-R values were based on underlying Sysmex-O values, explaining the similarity in analyzed data. In 5 cases where Sysmex-I was selected as Sysmex-R, Sysmex-I values showed excellent correlation to IRM.
No significant differences were found in the analysis of Oncology/Infection subcategory except that the 95% LA in Oncology patients was least in LH-750 (-9.178 to +5.723), significantly lesser than in the general analysis.
Positive predictive value and negative predictive value for over transfusion and under transfusion
A positive event for any method was a platelet count value of less than 10 10 9 /l and 20 10 9 /l respectively, at the two thresholds. The PPV for Sysmex-I was 79% at both the 10 10 9 /l and 20 10 9 /l thresholds [Table 2] suggesting that the Sysmex-I platelet counts when compared with the IRM were truly less than 10 10 9 /l and 20 10 9 /l respectively in 79% cases only. Rest 21% cases indicate the possible risk of over transfusion as the Sysmex-I on an average under estimates platelet counts. | Table 2: Analysis of Sysmex-R, Sysmex-I, Sysmex-O and LH-750 methods with the IRM at different transfusion thresholds
Click here to view |
Similarly, the risk of over transfusion for LH-750 was 15% and 21%, Sysmex-R is 10% and 7% at the two thresholds, whereas for Sysmex-O method the risk of over transfusion was the least at 5% and 7% respectively.
The risk of under transfusion at the threshold of 10 10 9 /l and 20 10 9 /l was least with LH-750 i.e., 3.4% and 1.4% respectively. However none of the instruments had an under transfusion risk of >10%.
The sensitivity and specificity at the two thresholds have also been calculated [Table 2].
| Discussion | | |
Historically, tests required for platelet enumeration have evolved from manual to semi-automated to completely automated systems.
The manual method that relies on a phase contrast microscope to count platelets is inefficient to say the least. It's bedeviled by issues such as being time intensive, subjective and a high degree on imprecision. The typical interobserver CVs is in the range of 10-25%. [8]
The advent of automated full blood counters have resulted in a dramatic improvement in precision, with typical CV's of <3%. [9] These counters use various methods for counting platelets. This includes aperture impedance, optical scattering and optical florescence.
Immunological flow reference method also known as the RBC platelet ratio method has replaced the manual phase contrast microscopy method as the reference method at the beginning of the last decade, i.e., 2000. [6]
Thus, we used this method to evaluate the platelet count accuracy of the Sysmex XE 2100 and Coulter LH-750 from patients with thrombocytopenia. Before comparing the analyzers to the IRM, precision studies to calculate the reproducibility were performed. The results are presented and compared with precision studies performed by Harrison et al. In [Table 3]. The reference method has a CV of 5.21%, which was quite similar to that found by Harrison et al. found. However, CV of LH-750 was lesser than 2%, which shows its high precision. The discrepancy in the CV results of LH-750 between the two studies could possibly be due to the different platelet counts taken for the study.
Our study showed a good correlation coefficient (>0.9) of all the automated hematology analyzers and IRM, which were similar to most studies. [11],[12],[13] However, "r" measures the strength of a relation between two variables, not the agreement between them, for which BA should be done. [7]
BA analysis showed that Sysmex-R had the least Bias and the narrowest 95%LA range and thus correlated best with IRM values. It is interesting to note that in as much as 95.7% of the cases, there was complete congruence between Sysmex-O and Sysmex-R figures, explaining the similarity in analyzed data. This suggests that the optical platelet count is superior to the impedance counts in thrombocytopenic patients. These results were similar to other recently published studies that suggest the superior accuracy of optical platelet counting methods for thrombocytopenic patients. [14],[15],[16],[17],[18] Sehgal et al. [11] suggested that the optical method of counting platelets had lesser bias and narrower 95% LA than impedance method; though, their study was performed on CellDyn sapphire. In our study, LH-750 had a strikingly similar 95% LA, which corroborates with the results of the recent studies that show impedance counts may show similar or even better results than optical platelet counting methods if curve fit correction is used. [12],[19] Infact, in the oncology subset, LH-750 had the narrowest 95% LA, which made it the best performer in this subgroup. Sehgal et al. [11] also showed that LH-750 had the least bias and the width of 95% LA were equally wide for LH-750 and Optical method.
Many studies have evaluated the degree of in accurate transfusion at different thresholds. [10],[13],[19],[20] In a study by Harrison et al. [10] the risk of over transfusion for LH-750, Sysmex-I, Sysmex-O and Sysmex-R at 10 10 9 /l was minimal as most of these methods in their study overestimated platelet counts, resulting in theoretical under transfusion of patients, rates of which varied from 5% to 11.7% [Table 4]. The most likely explanation for these findings is that all the patients included were taking cytotoxic chemotherapy and the debris due to therapy effect could be contributing to the overestimation of the counts despite the modern counting technologies used. Hong et al. [13] observed that at a threshold of 10 10 9 /l over transfusion ranged from 11.7% to 29.4% and under transfusion ranged from 11.7% to 23.5% with LH-750, Sysmex-I and Sysmex-O. However at a higher threshold of 20 10 9 /l, over transfusion ranged from 4.3% to 14.9% and under transfusion was 6.4% with all the methods. | Table 4: Comparison of over transfusion and under transfusion rates with other studies in the literature
Click here to view |
In our study, the risk of over transfusion was maximum (21%) with Sysmex-I at both the 10 10 9 /l and 20 10 9 /l thresholds, followed by LH-750 at 15% and 21%, Sysmex-R at 10% and 7% at the two thresholds, whereas for Sysmex-O method, the risk of over transfusion was the least at 5% and 7%, respectively. The risk of under transfusion at the threshold of 10 10 9 /l and 20 10 9 /l was least with LH-750 i.e., 3.4% and 1.4% respectively. However none of the instruments had an under transfusion risk of >10%.
The threshold level of 20 10 9 /l for blood transfusion is the most widely used for uncompromised patients. Recent studies have suggested that 10 10 9 /l or even 5 10 9 /l can be considered a new threshold for the uncompromised patients. [13] It has been shown that platelet transfusion thresholds of 10 10 9 /l as opposed to 20 10 9 /l reduce platelet transfusions by approximately 20% in leukemic patients without increasing any major risk of bleeding. [21] The lower threshold may have advantages; though, it should be viewed as a general reference value and not an absolute determinant to guide management. Clinical evaluation and careful patient monitoring must remain the prime factors in platelet transfusion therapy. In our study, for the thresholds of 10 10 9 /l or 20 10 9 /l, the proportion of under transfusion was low and none of the automated analyzers exceeded 10%. Over transfusion was found to be slightly more frequent than under transfusion. Thus, in addition to the clinical judgment, clinicians can feel quite confident about not transfusing a patient if the platelet count given by the automated analyzer is accurate and above the threshold level.
An important thing to remember is that in Sysmex-2100 overriding the "switching" algorithm and setting the analyzer to report the optical count at counts below a preset level, typically 50 10 9 /l should not be done as the XE 2100 gives the most accurate platelet count when the analyzer is allowed to use the "switching" algorithm to report the most accurate count. In a number of cases, the optical counts would be falsely high and so possibly result in patients being denied platelet transfusions when they are indicated.
Based on the current threshold levels, we conclude that the platelet counts determined by the LH-750 and Sysmex-R showed good correlation with the platelet counts as determined by the IRM. As already mentioned that LH-750 uses impedance with platelet fitting curve and Sysmex-R takes into account both optical and impedance, thus suggesting that both methodologies are good, provided we use them appropriately. In oncology patients, LH-750 proved to be better, suggesting that optical method may not be always required for platelet counting in oncology patients. However, it may be required in other categories, like infections and ITP where the size of the platelet may be large. Optical platelet count comes at an extra cost as one has to run the sample in the reticulocyte mode; however, it is worth it as unwarranted platelet transfusions can be avoided.
In this study, we have used the international reference method as the gold standard because of its proven analytical superiority. However, studies using IRM as the method of choice for predicting transfusion thresholds are still to be evaluated.
| Acknowledgment | | |
We would like to acknowledge National Health and Education Society for funding the project and Hinduja Hospital, where the entire work has been conducted.
| References | | |
| 1. | Ault KA. Platelet counting: Is there room for improvement? Lab Haematol 1996;2:139-43. 
|
| 2. | Ancliff PJ, Machin SJ. Trigger factors for prophylactic platelet transfusion. Blood Rev 1998;12:234-8.
[PUBMED] |
| 3. | Lawrence JB, Yomtovian RA, Dillman C, Masarik SR, Chongkolwatana V, Creger RJ, et al. Reliability of automated platelet counts: Comparison with manual method and utility for prediction of clinical bleeding. Am J Hematol 1995;48:244-50.
[PUBMED] |
| 4. | Lawrence JB, Yomtovian RA, Hammons T, Masarik SR, Chongkolwatana V, Creger RJ, et al. Lowering the prophylactic platelet transfusion threshold: A prospective analysis. Leuk Lymphoma 2001;41:67-76.
[PUBMED] |
| 5. | Schiffer CA, Anderson KC, Bennett CL, Bernstein S, Elting LS, Goldsmith M, et al. Platelet transfusion for patients with cancer: Clinical practice guidelines of the American Society of Clinical Oncology. J Clin Oncol 2001;19:1519-38.
[PUBMED] |
| 6. | International Council for Standardization in Haematology Expert Panel on Cytometry, International Society of Laboratory Hematology Task Force on Platelet Counting. Platelet counting by the RBC/platelet ratio method. A reference method. Am J Clin Pathol 2001;115:460-4.
|
| 7. | Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307-10.
[PUBMED] |
| 8. | Harrison P, Horton A, Grant D, Briggs C, MacHin S. Immunoplatelet counting: A proposed new reference procedure. Br J Haematol 2000;108:228-35.
[PUBMED] |
| 9. | Bentley SA, Johnson A, Bishop CA. A parallel evaluation of four automated hematology analyzers. Am J Clin Pathol 1993;100:626-32.
[PUBMED] |
| 10. | Harrison P, Segal H, Briggs C, Murphy M, Machin S. Impact of immunological platelet counting (by the platelet/RBC ratio) on haematological practice. Cytometry B Clin Cytom 2005;67:1-5.
[PUBMED] |
| 11. | Sehgal K, Badrinath Y, Tembhare P, Subramanian PG, Talole S, Kumar A, et al. Comparison of platelet counts by CellDyn Sapphire (Abbot Diagnostics), LH750 (Beckman Coulter), ReaPanThrombo immunoplatelet method (ReaMetrix), and the international flow reference method, in thrombocytopenic blood samples. Cytometry B Clin Cytom 2010;78:279-85.
[PUBMED] |
| 12. | Sandhaus LM, Osei ES, Agrawal NN, Dillman CA, Meyerson HJ. Platelet counting by the coulter LH 750, sysmex XE 2100, and advia 120: A comparative analysis using the RBC/platelet ratio reference method. Am J Clin Pathol 2002;118:235-41.
[PUBMED] |
| 13. | Hong KH, Kim MJ, Lee KW, Park KU, Kim HS, Song J. Platelet count evaluation using three automated haematology analysers compared with the immunoplatelet reference method, and estimation of possible inadequate platelet transfusion. Int J Lab Hematol 2009;31:298-306.
[PUBMED] |
| 14. | Briggs C, Harrison P, Machin SJ. Continuing developments with the automated platelet count. Int J Lab Hematol 2007;29:77-91.
[PUBMED] |
| 15. | Chapman DH, Hardin J, Miers M, Moyle S, Kinney MC. Reduction of the platelet review rate using the two-dimensional platelet method. Am J Clin Pathol 2001;115:894-8.
[PUBMED] |
| 16. | Kunicka JE, Fischer G, Murphy J, Zelmanovic D. Improved platelet counting using two-dimensional laser light scatter. Am JClin Pathol 2000;114:283-9.
|
| 17. | Stanworth SJ, Denton K, Monteath J, Patton WN. Automated counting of platelets on the Bayer ADVIA 120 analyser. Clin Lab Haematol 1999;21:113-7.
[PUBMED] |
| 18. | Briggs C, Harrison P, Grant D, Staves J, MacHin SJ. New quantitative parameters on a recently introduced automated blood cell counter - The XE 2100. Clin Lab Haematol 2000;22:345-50.
[PUBMED] |
| 19. | Segal HC, Briggs C, Kunka S, Casbard A, Harrison P, Machin SJ, et al. Accuracy of platelet counting haematology analysers in severe thrombocytopenia and potential impact on platelet transfusion. Br J Haematol 2005;128:520-5.
|
| 20. | Nishiyama M, Hayashi S, Futsukaichi Y, Suehisa E, Kurata Y. Usefulness of immunoplatelet measurement using the flow cytometric method for accurate platelet counting in hematological patients with severe thrombocytopenia. Rinsho Byori 2004;52:103-8.
[PUBMED] |
| 21. | Rebulla P, Finazzi G, Marangoni F, Avvisati G, Gugliotta L, Tognoni G, et al. The threshold for prophylactic platelet transfusions in adults with acute myeloid leukemia. Gruppo Italiano Malattie Ematologiche Maligne dell'Adulto. N Engl J Med 1997;337:1870-5.
|
Correspondence Address:
Tina Dadu
Department of Hematology, P. D. Hinduja Hospital & Medical Research Centre, Veer Savarkar Marg, Mahim, Mumbai - 400 016, Maharashtra
India
 Source of Support: National Health & Education Society (work
done at P.D. Hinduja Hospital & Medical Research Centre, Mumbai), Conflict of Interest: None  | Check |
DOI: 10.4103/0377-4929.118701
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3], [Table 4] |
|
| This article has been cited by | | 1 |
Alterations due to dilution and anticoagulant effects in hematologic analysis of rodent blood samples on the Sysmex XT-2000iV |
|
| Kaitlin A. Moorhead,Marielle L. Discipulo,Jing Hu,Roberta C. Moorhead,Jennifer L. Johns | | Veterinary Clinical Pathology. 2016; 45(2): 215 | | [Pubmed] | [DOI] | | | 2 |
Accuracy of a New Platelet Count System (PLT-F) Depends on the Staining Property of Its Reagents |
|
| Atsushi Wada,Yuri Takagi,Mari Kono,Takashi Morikawa,Gualtiero I. Colombo | | PLOS ONE. 2015; 10(10): e0141311 | | [Pubmed] | [DOI] | | | 3 |
Valutazione del sistema Cellavision DM-96 nell’analisi della formula leucocitaria e nel conteggio piastrinico |
|
| Maurizio Tala,Oriana Paoletti,Maria Annunciata Medagliani,Rossella Morandini,Anke Zimmermann,Claudia Dellanoce,Emilia Cancellieri,Sophie Testa | | La Rivista Italiana della Medicina di Laboratorio - Italian Journal of Laboratory Medicine. 2015; 11(2): 111 | | [Pubmed] | [DOI] | |
|
|
|
|
|
|
|
|
|
|
|
|
| Article Access Statistics | | | Viewed | 7999 | | | Printed | 256 | | | Emailed | 5 | | | PDF Downloaded | 232 | | | Comments | [Add] | | | Cited by others | 3 | | |
|
|
