Standardization of light transmittance aggregometry for monitoring antiplatelet therapy: an adjustment for platelet count is not necessary

Authors


Birgit Linnemann, J. W. Goethe University Hospital Frankfurt/Main, Division of Vascular Medicine, Department of Internal Medicine, Theodor-Stern-Kai 7, D-60590 Frankfurt/Main, Germany.
Tel.: +49 69 6301 5096; fax: +49 69 6301 7219.
E-mail: birgit.linnemann@kgu.de

Abstract

Summary. Background: Light transmittance aggregometry (LTA) is considered to be the ‘gold standard’ of platelet function testing. As LTA has been poorly standardized, we analyzed the results of LTA in healthy subjects and patients with antiplatelet therapy using different concentrations of agonists and performing tests in non-adjusted and platelet count-adjusted platelet-rich plasma (PRP). Methods: LTA was performed in 20 healthy subjects and in patients treated with aspirin (n = 30) or clopidogrel (n = 30) monotherapy, as well as in patients on combination therapy (n = 20), using arachidonic acid (ARA 0.25 and 0.5 mg mL−1) and adenosine diphosphate (ADP 2 and 5 μm) as agonists and performing platelet function tests in non-adjusted and platelet count (250 nL−1 ± 10%)-adjusted PRP. Results: The overall platelet aggregation response is decreased after adjusting the PRP for platelet count compared with measurements in unadjusted PRP. The variability of aggregation results is high in adjusted PRP in the subgroup of healthy subjects, ranging from 9.2–95.3% (5th–95th percentile) relative to 77.6–95.5% in non-adjusted PRP when determining maximum aggregation to ARA 0.5 mg mL−1. Late aggregation using ADP 2 μm ranges from 3.8–89.9% in adjusted PRP compared with 42.9–92.5% in non-adjusted PRP. Maximum aggregation using ARA 0.5 mg mL−1 in non-adjusted PRP differentiates between aspirin-treated patients and healthy controls well, whereas late aggregation using ADP 2 μm in non-adjusted PRP offers the best discrimination between clopidogrel-treated patients and healthy controls. Conclusion: Adjustment of PRP for platelet count does not provide any advantage and therefore the time-consuming process of platelet count adjustment is not necessary.

Introduction

Since its first description in 1962 by Born, platelet aggregometry has become the most widely used laboratory method to screen patients for inherited or acquired defects of platelet function [1,2]. Optical aggregometry measures the increase in light transmission through platelet-rich plasma that occurs when platelets are aggregated by adding an agonist. Although light transmittance aggregometry is suggested to be the gold standard for testing platelet function [3], there is still no international standardization of this technique. A panel of different agonists has been introduced, including arachidonic acid (ARA), epinephrine, adenosine diphosphate (ADP), collagen and ristocetin, and different optical aggregometers are in use [4]. However, the concentrations of agonists used for testing vary widely and accepted guidelines on how laboratories should test platelet function are lacking [5].

There are additional preanalytical and analytical variables that affect the results of platelet aggregation. For decades, it has been propagated that the so-called butterfly cannulae systems should be avoided for blood sampling because they would lead to considerable platelet activation [6]. In a recently published trial by our study group it was shown that the careful use of butterfly systems does not affect the results of optical aggregometry [7]. Furthermore, there is still controversy concerning whether the platelet count should be adjusted by mixing a subject’s platelet-rich plasma (PRP) with autologous platelet-poor plasma (PPP) to achieve platelet counts between 200 and 300 nL−1 [6,8]. On the one hand, it has been argued that in vitro aggregation is basically influenced by the platelet count in PRP and adjusting the platelet count has been widely recommended for platelet aggregation studies. On the other hand, it should be kept in mind that the mixing procedure itself could influence aggregation results. PPP may contain substances affecting platelet function that are released by platelets or other blood cells during centrifugation of blood samples, which is necessary to obtain PPP [6,9]. Therefore, using non-adjusted PRP may mean less manipulation of platelets and may represent, to a greater extent, the in vivo situation because the platelet count in native PRP correlates with blood platelet count.

Several parameters can be obtained from the aggregation curves of LTA. Only recently, the discussion has arisen over whether late aggregation is superior to maximum aggregation in ADP-induced LTA for evaluation of the responsiveness to clopidogrel medication. Up to now, there is no agreement on which parameter is the most sensitive to describe the efficacy of antithrombotic drugs.

Therefore, the aim of our study was to perform LTA in healthy volunteers and in aspirin and/or clopidogrel-treated patients with stable atherosclerotic disease using arachidonic acid and ADP in different concentrations as agonists and performing analysis in non-adjusted vs. platelet count-adjusted PRP in order to identify the conditions that offer the best discrimination between healthy subjects and patients receiving antithrombotic therapy.

Methods

Data sources and patients

Twenty healthy volunteers (nine males and 11 females; age 35.2 ± 12.1 years) and 80 patients on antithrombotic therapy for manifest atherosclerotic disease, such as ischemic heart disease, cerebrovascular disease or peripheral arterial disease (40 males and 40 females, age 66.5 ± 11.0 years), were included in this study. Among those on antithrombotic therapy were 30 patients receiving aspirin (100 mg day−1) and 30 patients receiving clopidogrel (75 mg day−1) as their only antithrombotic drug and a further 20 patients on combination therapy with aspirin (100 mg day−1) and clopidogrel (75 mg day−1). All patients were on stable antithrombotic medication for at least 4 weeks and stated that they had taken aspirin and/or clopidogrel within the last 24 h. All patients answered a standardized questionnaire assessing concomitant diseases and co-medication. Patients with a history of bleeding disorders or patients taking drugs other than aspirin or clopidogrel that are known to interfere with platelet function [e.g. non-steroidal anti-inflammatory drugs (NSAID)] within 2 weeks previous to blood sampling were not included in this study. The study was approved by the local ethics committee and written informed consent was obtained from all subjects.

Blood sampling

Blood was drawn by clean venipuncture from an antecubital vein using a 21-gauge butterfly cannula system (Multifly®-Set, 21 G × 1½ TW, 0.8 × 19 mm; Sarstedt, Nümbrecht, Germany). EDTA and citrate (0.106 mol L−1 trisodium citrate)-supplemented blood was collected into plastic syringes (Monovette®; Sarstedt). We ensured that the samples were mixed adequately by gently inverting the tubes. Platelet count was measured on the Sysmex® KX-21, an automatic multi-parameter blood cell counter. Platelet counts between 100 and 500 nL−1 were necessitated for subsequent platelet function testing. PRP was obtained by centrifuging citrated whole blood at room temperature at 140 × g for 5 min. To produce PPP (platelet count <10 nL−1), citrated whole blood was centrifuged more vigorously at 1500 × g for 15 min. Adjusted PRP with a platelet count of 250 nL−1 (± 10%) was obtained by diluting native PRP with the subject’s PPP. The time interval between blood sampling and testing was at least 1 h and did not exceed 3 h.

Light transmittance aggregometry

Platelet function tests were performed on the Behring Coagulation Timer® (BCT®; Dade Behring, Düdingen, Switzerland). The BCT® is a fully automatic machine used for routine and special coagulation testing. The BCT® detects platelet aggregate formation in PRP by changes in light transmission (monochromatic light, wavelength: 620 nm) at 37 °C. Platelet aggregation agonists (15 μL reagent) are introduced automatically to PRP (135 μL plasma) stirred at a velocity of 600 × g. In this study, the agonists adenosine 5-diphosphate (ADP; AppliChem, Darmstadt, Germany) and arachidonic acid (ARA; Moelab, Hilden, Germany) were used for stimulation of aggregation. The final agonist concentrations in PRP were 2 and 5 μm ADP and 0.25 and 0.5 mg mL−1 ARA. The extent of induced aggregation was defined by the slope of the aggregation curve obtained from the change in light transmission over time. An example aggregation recording is shown in Fig. 1. The maximum aggregation response, which is seen approximately 90 s after the addition of the agonists, was registered. Light transmission was measured in PRP at the start and at the time of the maximum aggregation and compared with PPP. Maximum aggregation was calculated after the formula:

image

We also recorded late aggregation in APD-induced aggregation after a period of 300 s as disaggregation is a frequent phenomenon, especially in clopidogrel-treated patients.

Figure 1.

 Aggregation curve of a clopidogrel-treated patient registered on the Behring Coagulation Timer (BCT®). The graph shows the change in light transmission (extinction mE) over time (s).

The 5th–95th percentile of each parameter measured in duplicate in the group of 20 healthy volunteers was considered as the reference range. This is in accordance with the actual recommendations of the International Society of Thrombosis and Haemostasis given at the 53rd Annual Scientific and Standardization Committee Meeting, Geneva 2007. Additionally, we determined the within-day precision of LTA by drawing a blood sample from five healthy volunteers, performing platelet function testing on five consecutive days and calculating the coefficients of variation (CV) of each parameter.

Statistics

Statistical analysis was performed using the Statistical Package for Social Sciences (SPSS version 12.0; SPSS, Chicago, IL, USA). Besides descriptive statistics comprising frequencies, mean, median, standard deviation, range and percentiles, correlation coefficients after Spearman were calculated. The criterion for statistical significance was a P-value less than 0.05. Coefficients of variation were calculated with the formula CV = (standard deviation/mean) × 100%. Results are also presented as box plots with the bare length indicating the interquartile range (25th–75th percentile). Outliers are defined as values differing 1.5–3.0 bare lengths, whereas extreme values are those differing >3.0 box lengths from the upper or lower edge of the box. In the figures, outliers are illustrated as circles and extreme values as stars.

Results

In the total cohort, platelet count ranged from 162 to 496 nL−1 in whole blood, from 318 to 976 nL−1 in non-adjusted platelet-rich plasma (PRP) and from 224 to 274 nL−1 in platelet count-adjusted PRP (Table 1). The correlation between platelet counts in whole blood and non-adjusted PRP according to Spearman’s correlation coefficient was high (= 0.78; < 0.001).

Table 1.   Baseline characteristics of healthy subjects and patients on antithrombotic medication
Patient groupAge (years)Sex (m = male, f = female)Platelet count (nL−1) ± SD
Whole bloodNon-adjusted PRPAdjusted PRP
Healthy volunteers (n = 20)35.2 ± 12.1m: n = 9 f: n = 11279.2 ± 60.0536.2 ± 104.1251.1 ± 12.7
Aspirin monotherapy (n = 30)64.5 ± 12.5m: n = 15 f: n = 15281.2 ± 67.0513.1 ± 130.2250.9 ± 9.4
Clopidogrel monotherapy (n = 30)69.9 ± 9.4m: n = 15 f: n = 15297.6 ± 85.4582.9 ± 158.8250.2 ± 11.6
Aspirin/clopidogrel combination therapy (n = 20)64.3 ± 10.2m: n = 10 f: n = 10254.5 ± 82.3476.6 ± 117.7251.3 ± 9.5

Healthy volunteers

With regard to the aggregation response of healthy volunteers (controls), we observed an aggregation of a lesser extent when using platelet count-adjusted relative to non-adjusted PRP, irrespective of which agonist was used (Figs 2 and 3). Moreover, the variability of the aggregation response in adjusted PRP was remarkably higher compared with non-adjusted PRP. In adjusted PRP, the range between the 5th and 95th percentile, which generally characterizes the reference range, is wide due to a respectable number of outliers and extreme values, especially when measuring maximum aggregation to ARA at the 0.5 mg mL−1 concentration and late aggregation to 2 μm ADP. This indicates that monitoring efficacy of antithrombotic therapy in adjusted PRP will not be suitable. In contrast, the coefficients of variation (CVs) determined from measurements in five healthy subjects on five consecutive days were less than 2.8% if non-adjusted PRP was used and less than 8.2% when measuring in platelet count-adjusted PRP, indicating the precision of the method. Reference ranges and CVs for each parameter are presented in Table 2.

Figure 2.

 Maximum aggregation with ARA 0.25 and 0.5 mg mL−1 as agonist comparing measurements in non-adjusted (dark grey) and platelet count-adjusted PRP (platelet count 250 nl−1± 10%) in healthy subjects and patients treated with aspirin or a combination therapy.

Figure 3.

 Maximum aggregation and late aggregation with ADP 2 and 5 μm as agonist comparing measurements in non-adjusted and platelet count-adjusted PRP (platelet count 250 nl−1 ± 10%) in healthy subjects and patients treated with aspirin, clopidogrel or a combination therapy.

Table 2.   Reference ranges (5th–95th percentile; Ref.) and coefficients of variation (CV) for maximum aggregation (MaxAgg) and late aggregation (LateAgg) determined by performing LTA in five healthy subjects on five consecutive days
Measurement in non-adjusted PRP
 MaxAggLateAgg
Ref. (%)CV (%)Ref. (%)CV (%)
ADP 2 μm56.2–91.41.6342.9–92.52.72
ADP 5 μm74.5–92.21.4072.1–93.01.47
ARA 0.25 mg mL−180.9–96.92.02
ARA 0.5 mg mL−177.6–95.52.76
Measurement in platelet count adjusted PRP
 MaxAggLateAgg
Ref. (%)CV (%)Ref. (%)CV (%)
ADP 2 μm33.4–85.35.963.8–89.98.12
ADP 5 μm52.7–87.72.9438.9–92.73.74
ARA 0.25 mg mL−14.1–95.42.54
ARA 0.5 mg mL−19.2–95.35.04

Aspirin-treated patients

Patients receiving aspirin as their only antithrombotic drug or in combination with clopidogrel showed good inhibition of platelet function using ARA as an agonist, irrespective of the agonist concentration (Fig. 2). Under combination therapy with clopidogrel, even more patients showed a good response in ARA-induced aggregation than under monotherapy with aspirin. When comparing maximum aggregation to ARA measured in adjusted vs. non-adjusted PRP, a better discrimination between normal aggregation in healthy subjects and impaired aggregation in aspirin-treated patients was obtained by performing LTA with non-adjusted PRP. ADP-induced aggregation was also impaired by aspirin, but to a lesser extent compared with patients treated with clopidogrel and the variability of late aggregation was high (Fig. 3).

Clopidogrel-treated patients

Patients treated with clopidogrel show a high variability of maximum and late aggregation responses to ADP at concentrations of 2 and 5 μm (Fig. 3). The variability is even larger for late relative to maximum aggregation. Comparing clopidogrel-treated patients with healthy subjects, late aggregation using ADP 2 μm in non-adjusted PRP provides a better discrimination than maximum aggregation although the correlation between these two parameters was high (= 0.92, < 0.001). Adjusting PRP for platelet count did not offer any advantage, but, if adjustment is performed, the 5 μm ADP concentration has to be used because healthy subjects without any antithrombotic medication show impaired platelet function to 2 μm ADP in adjusted PRP.

Comparison of platelet function in non-adjusted vs. platelet count-adjusted PRP

The scatter plots in Fig. 4 illustrate the correlation of the aggregation response in patients on antithombotic therapy measured in non-adjusted vs. adjusted PRP. In a respectable number of patients, high residual platelet activity despite antithrombotic medication in non-adjusted PRP contrasts with low residual platelet activity in adjusted PRP, and vice versa. This means that testing platelet function in non-adjusted and platelet count-adjusted PRP does not identify the same patients who would be considered as bad responders to aspirin or clopidogrel treatment. When performing LTA with ARA 0.5 mg mL−1, we observed residual platelet function with aggregation values in the reference range in 5/30 (16.7%) aspirin-treated patients if non-adjusted PRP was used as compared with 19/30 patients (63.3%) when adjusted PRP was used. In the determination of the late aggregation response to ADP 2 μm, we observed residual platelet function in 4/30 (13.3%), whereas corresponding values in platelet count-adjusted PRP could not be presented because the wide reference range is not convenient, as mentioned above. Regarding late aggregation using ADP 5 μm, we identified 3/30 (10.0%) and 7/30 (23.3%) patients with residual platelet function in non-adjusted and adjusted PRP, respectively. In addition, we observed no impact of platelet count in non-adjusted PRP on the aggregation response to ARA or ADP, neither in healthy subjects nor in aspirin- or clopidogrel-treated patients (< 0.2, = ns).

Figure 4.

 Correlation of aggregation response in non-adjusted vs. platelet count-adjusted PRP measuring ARA-induced aggregation in patients treated with aspirin (ASS; n = 30) and ADP-induced aggregation in patients treated with clopidogrel (CLOP; n = 30) as well as both in patients on combination therapy (ASS + CLOP; n = 20).

Discussion

When measuring platelet function by light transmittance aggregometry we observed that the overall platelet aggregation response was decreased after adjusting PRP for platelet count compared with measurements in unadjusted PRP. We did not observe that adjustment for platelet count would improve the results of platelet function testing as there was no association of platelet count in whole blood or PRP with the aggregation response to ARA or ADP measured in non-adjusted PRP. This could be shown in healthy subjects as well as in patients treated with aspirin or clopidogrel. Cattaneo et al. recently reported similar results in an investigation of ADP-induced aggregation in 151 healthy subjects [9]. The authors speculate that the decreased aggregation response in adjusted PRP might be due to release of substances from blood cells during more vigorous centrifugation of blood performed to obtain PPP. Additionally, we observed increased variability of the aggregation response when adjusting PRP for platelet count in the subgroup of healthy subjects, leading to a wide range between the 5th and 95th percentile, which is considered as the reference range in our laboratory according to the actual recommendations of the International Society of Thrombosis and Haemostasis presented at the 53rd Annual Scientific and Standardization Committee Meeting (Geneva 2007). In contrast, the between-day variation determined from platelet function tests performed in five healthy subjects on five consecutive days is <2.8% in non-adjusted and <8.2% in platelet count-adjusted PRP, indicating the high precision of LTA. Furthermore, platelet function testing in non-adjusted and adjusted PRP does not necessarily identify the same patients considered to be bad responders to either aspirin or clopidogrel. Therefore, non-adjusted PRP should be preferred for platelet aggregation studies, and the time-consuming process of platelet count adjustment is not needed, as it does not provide any advantage.

In addition, we found that late aggregation response to 2 μm ADP measured in non-adjusted PRP differentiates best between healthy controls and patients under clopidogrel treatment, whereas using the same concentration in adjusted PRP is not suitable for determining responsiveness to clopidogrel due to the high variability of results obtained from healthy subjects. It is remarkable that the best differentiation is obtained when using ADP at a concentration that is considered to be almost physiological (i.e. 1–2 μmol L−1). Many laboratories use higher concentrations of ADP for the induction of platelet aggregation and, as our analysis indicates, adequate measurements with 5 μm ADP can be performed in non-adjusted and platelet count-adjusted PRP. For aspirin-treated patients, discrimination between patients with impaired platelet function due to aspirin and healthy subjects was also superior when determining ARA-induced maximum aggregation in non-adjusted PRP. We did not observe any relevant difference between the ARA concentrations of 0.25 and 0.5 mg mL−1.

The measurement of late aggregation after a fixed period of time (i.e. 300 s after the addition of ADP) instead of maximum aggregation for evaluation of the efficacy of clopidogrel treatment by ADP-induced aggregation has been recommended by other authors recently [10,11]. It has been hypothesized that initiation of aggregation and the maximum aggregation response to ADP is mediated mainly by the P2Y1 receptor. The function of the P2Y12 receptor corresponds better to late aggregation as it is supposed to be responsible for the stabilization of platelet aggregates [12]. As the active metabolite of clopidogrel selectively inhibits ADP binding to P2Y12 but not P2Y1, it would be more convenient to monitor the inhibitory effects of clopidogrel therapy by measuring the late aggregation response. On the other hand, we found a strong correlation between maximum and late aggregation, which has been previously described [11]. In addition, Gurbel et al. found that both measurements were equivalent for determination of the incidence of non-responsiveness to clopidogrel [8]. This indicates that maximum and late aggregation might be interchangeable. However, clinical endpoint studies are necessary to determine which parameter is the better predictor of future cardiovascular events.

Although considered to be the traditional ‘gold standard’ of platelet function testing, light transmittance aggregometry remains poorly standardized and differs from laboratory to laboratory with regard to the panel of agonists, agonist concentration, the question of whether to adjust or not adjust PRP to a certain standardized platelet count, the type of aggregometer and analysis of the obtained aggregation curve. There is agreement that LTA is performed under non-physiological circumstances, as separated platelets are stirred under low shear stress conditions and only form aggregates after addition of agonists, conditions which do not simulate platelet adhesion, activation and aggregation upon endothelial damage of blood vessels. In response to the problems with LTA, a number of alternative tests have been developed that attempt to mimic primary hemostasis in vitro, including impedance whole blood aggregometry and fully automated cartridge-based systems such as the Platelet Function Analyzer-100® or the VerifyNow®. As the different tests show poor agreement and correlation between themselves and the results of LTA [13–16], the question remains whether these tests will be clinically useful in the prediction of future vascular thrombotic events. However, there have been clinical studies indicating that the results of LTA predict patient outcome [17,18]. Only recently, Buonamici et al. demonstrated that cardiovascular mortality was higher in patients who were considered to be low-responders to aspirin and clopidogrel in LTA after percutaneous coronary intervention (PCI) with drug eluting-stents [19]. So, despite the problems with LTA, this method remains suitable for monitoring antiplatelet drug therapy until there is clear evidence that other tests are superior at predicting recurrent vascular events. Therefore, laboratories should expend more effort in standardizing the method so that results from one laboratory can be transmitted to others.

We conclude that adjustment of PRP for platelet counts does not provide any advantage and, therefore, the time-consuming process of platelet count adjustment is dispensable. Furthermore, we demonstrated that maximum aggregation using ARA at a concentration of 0.5 mg mL−1 in non-adjusted PRP differentiates well between ASS responsive patients and healthy controls. Late aggregation response to ADP 2 μm using non-adjusted PRP offers the best discrimination between patients on clopidogrel treatment and healthy controls.

Disclosure of Conflict of Interests

The authors state that they have no conflict of interest.

Ancillary