Abstract
The objective of this study was to comparatively evaluate three commercial whole-blood platelet function analyzer systems: Platelet Function Analyzer-200 (PFA; Siemens Canada, Mississauga, Ontario, Canada), Multiplate analyzer (MP; Roche Diagnostics International Ltd., Rotkreuz, Switzerland), and Plateletworks Combo-25 kit (PLW; Helena Laboratories, Beaumont, TX, USA). Venipuncture was performed on 160 patients who visited a department of cardiology. Pairwise agreement among the three platelet function assays was assessed using Cohen’s kappa coefficient and percent agreement within the reference limit. Kappa values with the same agonists were poor between PFA-collagen (COL; agonist)/adenosine diphosphate (ADP) and MP-ADP (–0.147), PFA-COL/ADP and PLW-ADP (0.089), MP-ADP and PLW-ADP (0.039), PFA-COL/ADP and MP-COL (–0.039), and between PFA-COL/ADP and PLW-COL (–0.067). Nonetheless, kappa values for the same assay principle with a different agonist were slightly higher between PFA-COL/ADP and PFA-COL/EPI (0.352), MP-ADP and MP-COL (0.235), and between PLW-ADP and PLW-COL (0.247). The range of percent agreement values was 38.7% to 73.8%. Therefore, various measurements of platelet function by more than one method were needed to obtain a reliable interpretation of platelet function considering low kappa coefficient and modest percent agreement rates among 3 different platelet function tests.
Introduction
Because platelets play a vital role in primary hemostasis, analysis of platelet function is a crucial component in the assessment of the hemostasis status. Most analysis of platelet function has been used for the diagnosis and management of patients suffered with bleeding problem [1]. In perioperative settings, platelet function tests are also needed for patients receiving antiplatelet medication, such as cyclooxygenase-1 inhibitors, adenosine diphosphate antagonists, and glycoprotein IIb/IIIa inhibitors [2]. However, platelet function tests are still labor intensive and time-consuming and require special equipment and experts of specialized laboratories, in contrast to coagulation tests which are more standardized and fully automated such as prothrombin time (PT) or activated partial thromboplastin time (aPTT) [3].
To overcome these challenges, more convenient point-of-care tests for assessing platelet function in whole blood have been developed, specifically analyzer systems that have been coded as Platelet Function Analyzer (PFA)-200 (Siemens Canada, Mississauga, Ontario, Canada), Multiplate (MP) analyzer (Roche Diagnostics International Ltd., Rotkreuz, Switzerland), and Plateletworks (PLW) kit (Helena Laboratories, Beaumont, TX, USA) [2, 5]. These tests assess platelet aggregation in response to agonists by different methods and may be used to evaluate congenital and acquired platelet function defects, monitor antiplatelet therapy for cardiac conditions, and assess the risk of a hemorrhage and the need for blood products prior to cardiac surgery [3, 7].
PFA is a point-of-care platelet function test designed to be a surrogate of in vitro bleeding time [3]. It measures the time to occlusion (CT, closure time) of blood flow through a collagen-coated membrane under high shear stress conditions in the presence of epinephrine or ADP [3, 8]. As an improved version of PFA-100, PFA-200 has now become more usable because it has shown promising results in a few clinical studies [5].
The MP analyzer is a whole-blood aggregometry device using electrical impedance, which has gradually become preferred over platelet-rich plasma (PRP)-based optical aggregometry as the standard test in some laboratories [7]. Similar to optical aggregometry, the MP analyzer measures platelet function by determining aggregation after addition of an agonist. In contrast, instead of aggregating freely in liquid phase, platelets aggregate on two electrodes within a disposable cuvette [3, 8]. The MP analyzer requires a smaller sample volume and less manipulation of platelets, thereby decreasing the risk of platelet activation prior to analysis [3, 9]. Furthermore, it offers the advantage of examining platelet function in a more physiological environment because cellular components are present [4, 10]. Each measurement unit consists of two pairs of electrodes, allowing for simultaneous duplicate measurements [2]. The principle of the MP analyzer is based on the fact that platelets get sticky upon activation, and therefore tend to adhere to and aggregate on metal sensor wires in the Multiplate test cell. This phenomenon allows researchers to connect the test cell to the instrument to record the electrical resistance between the sensor wires during the test. The induced change in the impedance of the electrical system is continuously recorded and plotted against time, calculating aggregation velocity, total aggregation, and the area under the curve (AUC) [2].
The PLW assay is a whole-blood test assessing integrin αIIbβ (GP IIb/IIIa)-mediated platelet aggregation that requires no sample preparation [2]. It is based on a simple principle of counting platelets [2]. By comparing platelet counts within a control ethylenediamine tetra-acetic acid (EDTA) tube and after aggregation with a platelet agonist within citrated tubes, which decreases the number of individual platelets available for counting, the percentage of aggregated platelets can be calculated [6, 11]. Different test tubes containing collagen, ADP, or arachidonic acid as platelet agonists are available for assessing the effect of antiplatelet therapy [2]. It is approved for use in clinical practice to monitor the use of the antiplatelet drugs aspirin and clopidogrel [2, 11].
The objective of this study was to comparatively evaluate the three tests that assess platelet function in whole blood using different analytical principles: PFA by mechanical aperture closure, MP by electrical impedance, and PLW by platelet counting.
Materials and methods
Study protocol
The study protocol was approved by the Institutional Review Board of the Korea University Guro Hospital, and written informed consent was obtained from all the participants before enrollment into the study. Peripheral blood was collected from 160 patients who visited the Department of Cardiology of Korea University Guro Hospital. These patients were consecutively recruited for preventing coronary arterial spasm and we excluded patients who have underlying atherosclerotic coronary diseases confirmed by coronary angiogram in this study. There were not patients taking an antiplatelet medicine. Only one patient took an angiotensin-converting enzyme inhibitor (ACEI) because Asian patients taking and ACEI experienced a side effect such as a dry cough.
Blood samples were collected in tubes containing 3.2% sodium-citrate (0.109 mol/l buffered sodium citrate, BD Vacutainer Systems, Franklin Lakes, NJ, USA) for PFA-200 testing and hirudin (415 mg/ml; Roche; Mannheim, Germany) for MP testing by venipuncture. Blood collection for PLW testing was performed as follows. A baseline sample of blood was collected in an EDTA tube (BD Vacutainer Systems, Franklin Lakes, NJ, USA), while agonist-induced platelets were collected in a 3.2% sodium citrate tube after aggregation with either ADP (20μM) or collagen (10μg). The addition of blood to the agonist tubes activates platelets, which adhere to the tube, thus eliminating them from the platelet count.
After collection, the tubes were gently inverted (three to six times) and special care was taken with the sample transport to avoid platelet agitation. The blood samples were allowed to remain at room temperature for 30 min, and then these samples were gently remixed prior to analysis.
Platelet counts were measured using an automated blood analyzer (Sysmex XE5000, Sysmex Corporation, Kobe, Japan) within 2 hours of collection.
Platelet function analyzer (PFA)-200®
PFA-200 measures platelet aggregation induced by shear stress. The time needed for occlusion of a hole in the cartridge by the thrombocyte plug is called closure time (PFA-CT) and this is the main study parameter here. Assessment on PFA (Siemens Healthcare Diagnostics, Marburg, Germany) was performed using 800μl of each citrated blood sample with cartridges containing collagen-adenosine diphosphate (COL-ADP) and collagen-epinephrine (COL-EPI)-coated membranes for platelet activation (PF-COL/ADP and PF-COL/EPI) [6]. The cartridges were allowed to warm up to room temperature prior to use. Each sample was analyzed in duplicate or in triplicate if a sample error was encountered in one or both of the first two replicates. If we found a sample error such as flow obstruction, we collected resample from patients and analyzed the resample by PFA-200. The mean CT values, in seconds, for each subject were used for analysis. The PFA-CT has a theoretical maximum of 300 sec. Any closure time greater than 300 sec is reported as 300 sec. Furthermore, compared to PFA-100, it measures two more parameters: initial flow rate (PFA-IFR) and total volume (PFA-TV). The assay was performed using the manufacturer’s guidelines.
Multiplate (MP) analyzer
Platelet aggregation was conducted on the whole-blood five-channel MP analyzer (software version 2.03, Roche Diagnostics) by impedance aggregometry. This system uses test cells that contain two parallel pairs of electrodes. The metal sensors are made of highly conductive silver-coated copper. Activated platelets adhere and aggregate on the metal sensor wires, resulting in a rise in electrical resistance, which is continuously measured for 6 min. The results are expressed in AUC (area under the curve = area under the aggregation curve) units (U) and are recorded against time. The AUC reflects overall platelet activity, which is affected by the height of the aggregation curve.
Assessment on the MP analyzer was performed on hirudinated blood samples. Agonists tested were ADP (adenosine-5′-diphosphate; final concentration, 6.5μmol/l) and COL (collagen; final concentration, 3.2μg/ml). All the reagents were supplied by Roche Diagnostics (Roche Diagnostics International Ltd., Rotkreuz, Switzerland).
Each agonist was prepared as instructed by the manufacturer by adding 1 ml of distilled water to lyophilized stock. Reconstituted agonists were stored in 200-μl aliquots at –20°C for a maximum of 28 days (ADP) or 4°C for a maximum of 7 days (COL) and were allowed to warm up to room temperature immediately prior to analysis.
Plateletworks
Plateletworks assay was the evaluating the function of platelet aggregation by comparing impedance platelet counts between a sample anticoagulated with EDTA (baseline count) and a sample in a tube containing a platelet agonist such as ADP or collagen. Baseline platelet counts from the EDTA-anticoagulated samples and platelet counts from the PLW assay agonist tubes containing either ADP or COL (PLW-ADP and PLW-COL, respectively) were analyzed using an automated hematology analyzer (Sysmex XE5000, Sysmex Corporation, Kobe, Japan) that uses the impedance method for platelet enumeration. The percent platelet aggregation was calculated using the following formula:
% aggregation = [(EDTA platelet count – agonist-induced platelet count) ÷ EDTA platelet count]×100
Mean percent aggregation (% Agg) values calculated by means of impedance-based platelet counts, were used for analysis.
Statistical analysis
All the data were recorded in Microsoft Office Excel Software. Mean±SD values for continuous variables or n (%) for categorical values were presented. Pairwise agreement among the three platelet function assays was assessed using Cohen’s kappa coefficient or percent agreement. Cohen’s kappa coefficient results were interpreted as follows: values ≤0 meant no agreement and 0.01–0.20 none to slight, 0.21–0.40 fair, 0.41–0.60 moderate, 0.61–0.80 substantial, and 0.81–1.00 denoted almost perfect agreement [12]. The percent agreement was directly interpreted as the percentage of data that showed no difference between the values that were assigned either within the reference range or not (categorical agreement). Data with P values < 0.05 were considered statistically significant. All statistical analyses were conducted in the SPSS for Windows software package, version 20.0 (SPSS Inc, Chicago, IL, USA).
Results
Descriptive characteristics and hematological parameters
Fifty-three male and 107 female subjects, aged 60.86±10.01 years (mean±1SD), were included (Table 1). Among the subjects, 115, 57, and 13 patients were diagnosed with hypertension, dyslipidemia, and diabetes, respectively. Forty-two patients were diagnosed with hypertension and dyslipidemia, while 10 patients were diagnosed with hypertension and diabetes. Two had a diagnosis of hypertension, diabetes, and dyslipidemia. The clinical characteristics of the patients are shown in Table 1. The platelet counts were 216±56×109/l (mean±1SD).
Demographics of the study population
Demographics of the study population
Continuous data are shown as mean±1SD, dichotomous data are shown as N (%). Abbreviations: BMI, body mass index; ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker.
For PFA, median (range) CT values, using COL–ADP and COL-EPI cartridges, was 96 (54–300) and 154 (75–300) sec, respectively (Table 2). For the MP analyzer, median (range) area under curve (Units) for ADP and COL was 68 (17–119) and 51 (2–101), respectively. For the PLW assay, median (range) percent aggregation for ADP and COL agonists was 88.8 (14.2–113) and 92.8 (11.5–98.8), respectively, using platelet counts measured by an impedance hematology analyzer.
Means, medians, and ranges for whole-blood platelet function testing among 160 subjects using three analyzer systems
Means, medians, and ranges for whole-blood platelet function testing among 160 subjects using three analyzer systems
Abbreviations: ADP = adenosine diphosphate; AUC (U) = area under curve screenshot Units [=AU×min/10 and rounded down]; % Agg = percent aggregation; COL = collagen; EPI = epinephrine; MP = Multiplate Analyzer (Roche Diagnostics International Ltd., Rotkreuz, Switzerland); PFA = Platelet Function Analyzer-200 (Siemens Canada, Mississauga, Ontario, Canada); PW = Plateletworks (Helena Laboratories, Beaumont, TX). *Sysmex XE5000, (Sysmex Corporation, Kobe, Japan). †Normally distributed data.
Table 3 presents Cohen’s kappa coefficients for all the test combinations of the three methods. Agreement rate between PFA-200 COL/ADP and PLW-COL was the highest (58.1%, kappa value –0.067), but agreement rate between PFA-200 COL/ADP and MP-COL was the lowest (38.7%) for the same agonist with different assay principles. Better agreement rates were noticed for different agonists with the same assay principle (46.9–73.8%). Kappa values for the same agonist with different assay principles were poor between PFA-COL/ADP and MP-ADP (–0.147), PFA-COL/ADP and PLW-ADP (0.089), MP-ADP and PLW-ADP (0.039), PFA-COL/ADP and MP-COL (–0.039), and between PFA-COL/ADP and PLW-COL (–0.067). Higher kappa values were observed for different agonists with the same assay principle (0.235–0.352). There were no significant differences between dyslipidemia (n = 57) and the other group (n = 103) according to the three platelet function assays.
Pairwise agreement rates among the three platelet function methods, using Cohen’s kappa values
Pairwise agreement rates among the three platelet function methods, using Cohen’s kappa values
Abbreviations: ADP = adenosine diphosphate; CI = confidence interval; COL = collagen; EPI = epinephrine; MP = Multiplate Analyzer (Roche Diagnostics International Ltd., Rotkreuz, Switzerland); PFA = Platelet Function Analyzer-200 (Siemens Canada, Mississauga, Ontario, Canada); PLW = Plateletworks (Helena Laboratories, Beaumont, TX); N/A = not applicable.
The three main axes of traditional platelet function tests are the optical aggregometry, whole blood aggregometry and lumiaggregometry, but they are invasive, time-consuming, cost-ineffective, and need skillful technicians and a large sample volume [13]. Accordingly, new options for platelet function tests, such as PFA-200, MP analyzer, and PLW, have been developed. Nonetheless, those tests are based on other methods and use different agonists (ADP, COL, and EPI), reflecting the multiple pathways of platelet activation and multiple targets of antiplatelet drugs [14]. The MP analyzer is a semi-automated impedance aggregometry, which offers a low-shear system, and specific detection of aspirin, anti-GPIIb/IIIa, and anti-P2Y12 is possible. PLW is based on platelet counting and can detect three types of antiplatelet drugs just as the MP analyzer can [2]. In the present study, the three platelet function tests based on different analytical principles were evaluated comparatively.
Kappa values were fair (0.24–0.35) in the comparisons involving the same platelet function instruments with different agonists, between PFA-COL/ADP and PFA-COL/EPI, MP-ADP and MP-COL, and PLW-ADP and PLW-COL. Kappa values were much lower in the comparisons involving the same agonists with different platelet function instruments. These results are in line with those of another study, in which kappa values ranged from –0.02 to 0.37 [15].
Categorical pairwise agreement rates among the three platelet function tests ranged from 38.7% to 62.8%. These results are lower than those of the above study, which evaluated PFA-100 in comparison with light transmittance aggregometry (74.4%), flow cytometry (63.3%), and MP (75.6%) [15] (Major concern 3).
The low kappa values and modest agreement rates may be partially attributed to variable agonist types and concentrations across the tests. Another reason may be the difference in anticoagulants. Anticoagulants represent a significant variable in platelet function testing. PFA and PLW were tested using freshly citrated whole blood, whereas the MP analyzer was tested using hirudinized whole blood. Different anticoagulants prevent blood clotting by different mechanisms, each creating unique conditions that can modify platelet function [10]. Other reports have shown that ADP- and COL-induced whole-blood platelet aggregation is significantly lower in blood samples anticoagulated with sodium citrate compared with hirudin [10, 16].
The use of PFA-100 with COL-ADP and COL-EPI has been reported frequently. In a study based on 12 whole-blood samples collected with BD vacutainer plastic whole-blood tubes containing 0.109 mol/l buffered sodium citrate, the mean COL-ADP CT was 90 sec, while the mean COL-EPI CT was 123 sec [17]. In contrast, the mean COL-ADP and COL-EPI CT was 106.5 and 180.9 sec, respectively, in this study. Although sample handling after collection was similar in the previous and current study, these differences may be due to other pre-analytical variables in sample collection [17]. Nevertheless, this discrepancy indicates that the subjects in this study have lower rates of platelet adhesion and aggregation [5].
Meanwhile, patients’ characteristics such as platelet count <100×109/l and hematocrit <30% usually result in prolongation of CT [18]. Based on these criteria, this study included only four and three patients with a low platelet count (4/160) and low hematocrit (3/160), thereby minimizing the bias.
In a study aimed at developing reference ranges for platelet aggregation using the MP analyzer, the medians of AUC values for ADP and COL agonists in hirudinized blood were 87.0 and 81.5, respectively. On the other hand, in the present study, for the same anticoagulant and the same agonists at the same concentrations, the medians of AUC values for ADP and COL agonists were 68.0 and 51.0, respectively. This means that the subjects in this study have weaker platelet aggregation than those in the previous study. The differences could be caused by not only ethnic differences but also the presence of disease groups in the current study [10]. In addition, depending on the studies, degrees of responses to COL and ADP have been different for the MP analyzer [7, 10]. In the present study, responses to ADP were higher than responses to COL. Further evaluation of the MP analyzer for aggregation responses according to agonists is needed to facilitate its clinical application.
PLW correlates well with the “gold standard” of platelet aggregometry [6] and is known to be acceptable in terms of its precision, accuracy, and linearity when compared with a central hospital laboratory analyzer such as the Beckman Coulter Unicel DXH 800 [19]. Normal human reference intervals of 80–100% and 70–100% for ADP and COL, respectively, are provided by the manufacturer. Nonetheless, one study suggested that median % of aggregation for ADP and COL agonist is 71 and 49, respectively, which are lower counts in comparison with the present study [4]. The reason for this discrepancy could be the differences in enrolled disease populations such as dyslipidemia in the present study, because in the earlier study, a healthy population was recruited. Nevertheless, the influence of dyslipidemia was not detected in any of the three platelet function assays with agonists ADP and collagen in this study. It is likely that less stable aggregates in healthy people may increase platelet counts in the PLW assay in ADP and COL tubes and thus underestimate % Agg [4].
This study has limitations. First, three platelet function tests were compared in the absence of a gold standard measurement, leaving us without an assessment of accuracy. The second limitation is related to sample handling. In this study, blood samples were re-suspended after a rest period, and this approach could have led to more platelet activation than with continuous rocking or prompt analysis [18].
Overall, the Cohen’s kappa values for all the different assay combinations of three platelet function tests are –0.147 to 0.098 and the range of percent agreements is 38.7–62.8%. In conclusion, a single test result on platelet function may be difficult to interpret because of the lack of strong agreement among platelet function tests, wide 95% confidence intervals, their unknown clinical relevance and limited experiences with the use of these instruments. Although various platelet function tests have been developed over the years, implementation of a platelet function test involving more than one method is needed for reliable interpretation [15]. If the results of platelet function tests are inconsistent with clinical manifestation, additional tests are recommended. Moreover, prior to the routine clinical use of these tests, further evaluation of pre-analytical variables (e.g., sample collection and handling) and analytical variables (e.g., anticoagulant and agonist concentrations) is also needed.
Declaration of interest statement
The authors report that they have no conflicts of interest.
Footnotes
Acknowledgments
This study was supported by a grant from the Korea Health Technology R&D Project, Ministry of Health & Welfare, Republic of Korea (Grant No. HI14C0670).