Thrombosis Research (2008) 123, 100107

www.elsevier.com/locate/thromres

REGULAR ARTICLE

Monitoring aspirin treatment in patients with thrombocytosis: Comparison of the platelet function analyzer (PFA)-100 with optical aggregometry
Argirios E. Tsantes a,, Georgios Mantzios a , Vassiliki Giannopoulou b , Panagiotis Tsirigotis b , Stefanos Bonovas c , Evdoxia Rapti a , Elissavet Mygiaki a , Zafirios Kartasis d , Nikolaos M. Sitaras c , John Dervenoulas b , Anthi Travlou a
Laboratory of Haematology & Blood Bank Unit, Attikon General Hospital, School of Medicine, University of Athens, Greece b 2nd Department of Internal Medicine, Attikon General Hospital, University of Athens, Greece c Department of Pharmacology, School of Medicine, University of Athens, Greece d Department of Haematology, General Hospital of Chalkida, Greece
a

Received 23 December 2007; received in revised form 5 March 2008; accepted 6 March 2008 Available online 21 April 2008

KEYWORDS
PFA-100; Aggregometry; Aspirin; Thrombocytosis

Abstract Aspirin provides satisfactory protection against thrombotic episodes in essential thrombocythemia (ET), but at higher platelet counts has been less effective. Our aim was to compare the platelet function analyzer (PFA)-100 with optical aggregometry in order to determine a reliable method in monitoring aspirin's influence on platelet function in patients with thrombocytosis. We studied 36 patients with thrombocytosis. Sixteen of them, receiving aspirin, composed group A, while group B consisted of 20 patients not taking aspirin. In all patients, we compared the platelet function measured by classic optical aggregation tests with closure times (CT) obtained by the PFA-100. The definition of platelet responses as normal or pathological showed that PFA-100 collagen and/or epinephrine (CEPI) CTs and epinephrine-induced aggregometry is the pair of methods with the higher agreement in monitoring of platelet dysfunction due to ASA treatment (a = 94%). Satisfactory results were also obtained for group B (a = 81%). The comparison between PFA-100 CEPI CTs and arachidonic acid-induced aggregometry exhibited moderate agreement both in the total number of patients

Corresponding author. Xanthippis 75 Str., GR 10444, Athens, Greece. Tel.: +30 2105154915. E-mail address: atsantes@yahoo.com (A.E. Tsantes). 0049-3848/$ - see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.thromres.2008.03.008

Evaluation of PFA-100 in monitoring aspirin treatment in thrombocytosis

101

and in group A (a = 79% and 94%, respectively). PFA-100 collagen and/or ADP (CADP) CTs and ADP-induced aggregometry were not concordant. The PFA-100 system appears to be a reliable and rapid method in the assessment of aspirin's antiplatelet effect in patients with thrombocytosis. Regarding aggregometry, the selection of the inducer, its concentration and cut-off points is crucial in defining the response to antiaggregating agents. It still remains to determine whether there is any relevance between the measurements obtained by these methods and clinical outcome in thrombocythemic patients. 2008 Elsevier Ltd. All rights reserved.

Introduction
In essential thrombocythemia (ET), a variable number of patients present with thrombotic or hemorrhagic problems [1]. Uncontrolled thrombocytosis, activation of platelets and leukocytes and their interaction to form aggregates, in addition to prothrombotic circulating and endothelial factors have been involved in the pathophysiology of thrombosis in ET [2]. The precise incidence of thrombosis is hard to ascertain, but it has been reported to be more common than hemorrhage, and arterial thrombi are more common than venous [3]. Further, injudicious use of aspirin and an acquired von Willebrand factor (VWF) deficiency in plasma at higher platelet counts may result in a spontaneous bleeding tendency [4]. On the other hand, patients with reactive thrombocytosis (RT) may have platelet (PLT) counts as high as those of patients with ET, but hemorrhage and thrombosis are unusual. Aspirin has an established role in prevention of vascular disease and is recommended for all patients with ET in the absence of active contraindications [2]. Due to the difficulty in predicting thrombotic or hemorrhagic complications in ET patients, it would be of great importance to use a reliable and easy method in order to evaluate platelet function and monitor aspirin therapy in such patients. The platelet function analyzer (PFA)-100 (Dade Behring, Marburg, Germany) is an instrument and test cartridge system in which the VWF and collagendependent platelet adhesion and subsequent platelet aggregation is measured [5,6]. The instrument estimates the ability of platelets activated in a highshear environment to occlude an aperture in a membrane treated with collagen and epinephrine (CEPI) or collagen and ADP (CADP). The time taken for flow across the membrane to stop (closure time, CT) is recorded. The PFA-100 may be a useful tool in evaluation of a variety of clinical situations. It is helpful in diagnosing VWF abnormalities [7], other congenital disorders of platelet function [8], acquired platelet dysfunction [9], and monitoring antiplatelet drug therapy with agents such as acetyl salicylic acid (ASA) [10]. However, PFA-100 is a global

test and therefore sensitive to large number of variables known to affect platelet function in vivo like platelet count, haematocrit (Hct), drug effects, major platelet receptor defects or VWF defects [11]. How the PFA-100 system works with specimens with platelet counts N 500,000/L or Hct levels N 50% has not been completely delineated. Our aim was to investigate whether PFA-100 is a reliable method in monitoring the effects of aspirin on platelet function in patients with thrombocytosis (platelet count N 500,000/L). The PFA-100 measurements were compared with light transmission agrregometry values induced by several agonists, since aggregometry is often perceived as reference method for the platelet function tests. Materials & Methods
Our study included the following groups of patients: Group A consisted of 10 patients with ET and 6 with reactive thrombocytosis (RT) all receiving aspirin (100 mg per day), in order to examine the effect of elevated platelet count on PFA-100 measurements. Group B serving as control, included 12 patients with ET and 8 with RT who had not taken aspirin or other antiplatelet drugs for at least 7 days before their testing. 17 of the ET patients were taking hydroxyurea, two of them were anagrelide-treated, one interferon- treated, and two did not receive any cytoreductive therapy. Diagnosis of ET was based on the revised criteria of the Polycythemia Vera Study Group [12]. In all patients with RT, ET had been excluded and the underlying cause had been determined. Patients with RT included five with malignancy, three with infections, three who had undergone major surgical procedures, two with chronic inflammation, and one who was post splenectomy. Blood samples were obtained after informed consent was given. Aggregation tests and PFA-100 measurements were performed on the same day and within two hours of sampling. The results obtained from both methods were compared between groups A and B, in order to estimate the reliability of these methods in detecting aspirin response in patients with thrombocytosis. Then, in all patients we compared the platelet function measured by classic optical aggregation tests with CT obtained by the PFA-100. Platelet optical aggregometry was selected as a reference method since it is a classic, simple and reliable way, although time-consuming and difficult, to evaluate platelet function in elevated platelets counts and define the response to antiaggregating agents [13,14]. Full blood counts were performed on Sysmex XE-2100 analyzer (Roche, IL, USA) and VWF antigen (VWF:Ag) and activity (VWF activity) were measured on the blood samples as described below. Since the PFA-100 is

102
Table 1 Descriptive statistics Group A (n = 16)
Age Sex (males %) Hct (%) WBC (/L) PLT ( 103/L) VWF Activity (%) VWF:Ag (%) VWF Activity/VWF:Ag 59.1 13.0, 60.5 (3882) 6/16, 37.5% 41.4 4.0, 42.1 (32.347.2) 7800 3130, 7200 (400016400) 722 204, 655 (5101200) 124.5 93.8, 89.5 (39.4391.6) 170.0 124.4, 128.7 (39.2521.2) 0.77 0.22, 0.76 (0.271.28)

A.E. Tsantes et al.

Group B (n = 20)
54.3 14.7, 48.5 (3177) 9/20, 45.0% 39.9 5.2, 39.2 (32.248.0) 8750 3150, 9050 (380017000) 652 142, 623 (500994) 131.2 105.4, 87.3 (30.0405.4) 189.1 147.0, 149.7 (4.1566.4) 0.75 0.11, 0.75 (0.541.00)

Group A + B (n = 36)
56.4 14.0, 55.0 (3182) 15/36, 41.7% 40.6 4.7, 40.0 (32.248.0) 8340 3140, 7950 (380017000) 683 174, 650 (5001200) 128.2 98.8, 87.3 (30.0405.4) 180.4 135.5, 148.3 (4.1566.4) 0.76 0.17, 0.75 (0.271.28)

Group A: Patients taking ASA, with PLT N 500,000/L. Group B: Patients not taking ASA, with PLT N 500,000/L. Group A + B: Patients with PLT N 500,000/L. Data are presented as mean SD, median and (range). influenced by platelet count, Hct, and VWF defects, the correlation of these parameters with the PFA-100 measurements were estimated. reservoir of 1 CADP and 1 CEPI cartridge (prewarmed to room temperature) and then loaded into the PFA-100. The references ranges were 85164 sec for CEPI and 71118 sec for CADP cartridge. A pathological platelet response was therefore defined as a CEPI CT N 164 seconds and a CADP CT N 118 seconds. Like aggregometry, the pathological platelet response was further classified as decreased or null if prolongation of CTwas more than 164 or 300 seconds for CEPI cartridge and 118 or 300 seconds for CADP cartridge, respectively.

Platelet aggregation
The whole blood specimen was collected in 3.8% trisodium citrate and centrifuged at 200 g for 10 minutes to obtain platelet-rich plasma (PRP). The remaining specimen was recentrifuged at 2000 g for 15 minutes to obtain platelet-poor plasma (PPP). The platelet count was adjusted to between 200,000/l and 300,000/l with PPP . Aggregation was performed using a Biodata-PAP-4 aggregometer (Bio/Data Corporation, PA, USA). The 100% line was set using PPP and a 0% baseline established with PRP before addition of the agonist. A panel of 3 different agonists was used: ADP 2.010 5 M, arachidonic acid (AA) 500 g/ML, and epinephrine (EPI) 1.010 4 M (Bio/Data Corporation, PA, USA). Test procedure: 0.45 mL PRP were transferred into a cuvette incubated at 37 C for 3 minutes. Then, 0.05 mL of the agonist was added into the PRP and the aggregation pattern was allowed to generate for 5 minutes. Typical normal platelet aggregation responses at 5 minutes were considered the following values: for ADP 6389%, for AA 6590%, and epinephrine 54101%. Platelet responses lower than these values were reported as pathological and were further classified as decreased or null if increases of transmittance were less than lower normal values or 10% respectively [15]. With AA, decreased or null response was defined as increases of transmittance less than 20% [14] or 10%, respectively.

VWF antigen and activity

An automated latex enhanced immunoassay (HaemosIL assay, Instrumentation Laboratory Company, Lexington, USA) for the quantitative determination of VWF antigen and activity in citrated plasma on IL (Instrumentation Laboratory) coagulation systems (ACL TOP) was performed. Particularly the activity of VWF was determined by measuring the increase of turbidity produced by the agglutination of the latex reagent [16]. A specific anti-VWF monoclonal antibody adsorbed onto the latex reagent, directed against the platelet biding site of VWF (glycoprotein Ib receptor), reacts with the VWF of patient plasma. The degree of agglutination is directly proportional to the activity of VWF in the sample and is determined by measuring the decrease of transmitted light caused by the aggregates. Results were reported in % of normality. Values less than 42% and 40.3% were considered abnormal for VWF antigen and activity,

PFA-100
Table 3
For this test, whole blood collected in 3.8% trisodium citrate. 0.8 ml of the mixed whole blood was pipetted into the sample PFA EPI -vs- AG EPI

Group A
a = 75%, k = 0.47, p = 0.018 a = 69%, k = 0.21, p = 0.07 a = 36%, k = 0.24, p = 0.83

Group B
a = 63%, k = 0.27, p = 0.07 a = 65%, k = 0.00, a = 75%, k = 0.14, p = 0.72

Group A + B
a = 69%, k = 0.53, p b 0.001 a = 67%, k = 0.48, p b 0.001 a = 57%, k = 0.05, p = 0.62

Table 2 Group A
PFA-100 CEPI CT (secs) AG EPI (%) AG AA (%) 261.4 58.3 11.1 13.0 1.8 3.9

PFA EPI -vs- AG AA

Group B
165.1 69.9 56.0 42.6 88.1 9.5 PFA ADP -vs- AG ADP

PFA-100 CEPI and platelet aggregation with EPI and AA results. Group A: Patients taking ASA, with PLT N 500,000/L. Group B: Patients not taking ASA, with PLT N 500,000/L. CT = closure time, AG = aggregometry, EPI = epinephrine, AA= arachidonic acid. Data are presented as mean SD.

Agreement between the different tests determined by kappa statistics and the respective p-values. We defined the responses as normal, decreased, or null (2 cutoff points). a = agreement between the two methods, k = kappa statistic, AG = aggregometry, EPI = epinephrine, AA= arachidonic acid.

Evaluation of PFA-100 in monitoring aspirin treatment in thrombocytosis

Table 4 Group A
PFA EPI -vs- AG EPI a = 94%, k = 0.00, a = 94%, k = 0.00, a = 36%, k = 0.24, p = 0.83

103

Group B
a = 81%, k = 0.59, p = 0.009 a = 65%, k = 0.00, a = 75%, k = 0.14, p = 0.72

Group A + B
a = 88%, k = 0.72, p b 0.001 a = 79%, k = 0.58, p b 0.001 a = 57%, k = 0.05, p = 0.62

PFA EPI -vs- AG AA

respective p-value. Spearman's rho of b 0.20 is considered to indicate very weak correlation, 0.21 to 0.40 indicates weak correlation, 0.41 to 0.60 indicates moderate correlation, 0.61 to 0.80 indicates strong correlation, and N 0.81 indicates very strong correlation. For all tests, a probability level of b 0.05 was considered to be statistically significant. Stata software was used for all statistical analyses (Stata Corp., College Station, TX, USA).

PFA ADP -vs- AG ADP

Results
Of the 36 patients studied, 15 were male and ages ranged from 31 to 82 years. The descriptive characteristics, haematological parameters and VWF antigen, activity and VWFActivity / VWF: Ag ratio of patients are presented in Table 1. All patients were analyzed by PFA100 and light transmission aggregometry. In group B, 2 measurements of aggregation with EPI and 2 measurements of aggregation induced by AA were not performed due to deficiency of the respective reagent. Flow obstruction occurred in 2 out of 36 samples in CEPI cartridge measurements and in 5 out of 36 samples in CADP cartridge tests. The total percentage for flow obstruction in our PFA-100 studies was 9.7%. The platelet counts of the patients whose sample testing with PFA-100 failed due to flow obstruction were not statistically significantly different than the counts of the rest patients. A statistically significant difference was found between aspirin users and non-users regarding PFA-100 CEPI CTs (p b 0.001), and aggregometry with EPI (p =0.002), AA (p b 0.001) and ADP (p =0.010). This finding supports the suitability of these methods in detecting aspirin response in patients with thrombocytosis. Moreover, in group B, no patient with ET exhibited abnormal PFA-100 CEPI CTs versus normal aggregation with EPI, while in two cases normal PFA-100 CT versus abnormal platelet aggregation induced by EPI was observed (PFA EPI CT: 155 vs AG EPI: 6 and PFA EPI CT: 144 vs AG EPI: 20). Table 2 shows the results for comparison of PFA-100 CEPI with platelet aggregation induced by EPI and AA in both groups. All the results concerning the comparisons between the two methods are presented in Table 3 (using 2 cut-off points) and Table 4 (using 1 cutoff point).

Agreement between the different tests determined by kappa statistics and the respective p-values. We defined the responses as normal or pathological (1 cut-off point). a = agreement between the two methods, k = kappa statistic, AG = aggregometry, EPI = epinephrine, AA= arachidonic acid.

respectively. When the VWF Activity / VWF:Ag ratio was low (b 0.7), qualitative defect may be diagnosed.

Statistical analysis
Descriptive statistics of the data are presented as meanSD, median and range. Comparison of the PFA-100 and optical aggregometry measurements between groups A and B were performed using the two-sample Wilcoxon rank-sum (Mann-Whitney) test. Agreement between the different tests was determined by kappa statistics and the respective p-values. Kappa values of b 0.20 are considered to indicate poor agreement, 0.21 to 0.40 indicate fair agreement, 0.41 to 0.60 indicate moderate agreement, 0.61 to 0.80 indicate good agreement, and N 0.81 indicate very good agreement. Correlations among the different variables (PFA-100, PLT, Hct, VWF Activity, VWF:Ag, VWF Activity / VWF-Ag) were calculated by the non-parametric Spearman rank correlation coefficient and the

Figure 1 Comparison of all samples (n = 32) tested by the PFA-100 CEPI cartridge and by optical aggregometry using epinephrine (EPI) 1.0 10 4 M as an agonist. PFA-100 CT N 164 sec or aggregation b 54% (indicated by the lines on the graph) was determined as pathological platelet response.

104
The classification of platelet response in normal, decreased, or null (2 cut-off points) in all patients yielded the following results: PFA-100 CEPI CTs and aggregometry with EPI agreed in 69% of the samples, and this agreement was statistically significant (k = 0.53, p b 0.001). In similar, PFA-100 CEPI CTs and aggregometry with AA were concordant in 67% of the samples and this agreement was statistically significantly greater than chance (k= 0.48, p b 0.001). On the contrary, PFA-100 CADP CTs and aggregometry with ADP were not concordant (a = 57%, k = 0.05, p = 0.62). When the responses were defined as normal or pathological (1 cut-off point), the overall agreement between PFA-100 CEPI CTs and aggregometry with EPI was good (a=88%, k=0.72, p b 0.001) (Fig. 1). Similarly, the agreement between PFA-100 CEPI CTs and AAinduced aggregometry was moderate (a=79%, k=0.58, p =0.001) (Fig. 2). On the contrary, PFA-100 CADP CTs and aggregometry with ADP were not concordant. Analysis of the data for all patients, using the non-parametric Spearman rank correlation coefficient (rho), revealed that both PFA-100 CEPI CT and CADP CT were inversely correlated to VWF Activity (rho = 0.37; p = 0.036 and rho = 0.48; p = 0.008, respectively). PLT count was also inversely correlated to VWF Activity / VWF: Ag (rho = 0.49, p = 0.004). Further, we performed the above-mentioned statistical analyses separately for patients treated with ASA (group A, n = 16), and for patients not taking ASA (group B, n = 20). In group A, when we classified the responses using 2 cut-off points, we observed that PFA-100 CEPI CTs and EPI-induced aggregometry agreed in 75% of the samples, and this agreement was statistically significantly greater than chance (k= 0.47, p = 0.018). PFA-100 CEPI CTs and aggregometry with AA were concordant in 69% of the samples. However, this agreement was marginally not statistically significantly greater than chance (k = 0.21, p = 0.07). By defining the responses with 1 cut-off point, the overall agreement between PFA-100 CEPI CTs and aggregometry with EPI was found very high (a = 94%), however a kappa statistic and a respective p-value could not be estimated. In similar, the agreement between PFA-100 CEPI CT and aggregometry with AA was very high (a = 94%).

A.E. Tsantes et al.

Analysis of the data, using the non-parametric Spearman rank correlation coefficient (rho), showed that PFA-100 CADP CT was inversely correlated to VWF Activity (rho = 0.62, p = 0.017) and PLT count was again inversely correlated to VWF Activity / VWF: Ag (rho = 0.54, p = 0.029). In group B, when we classified the responses as normal, decreased, or null (2 cut-off points), we observed that PFA-100 CEPI CTs and aggregometry with EPI agreed in 63% of the samples, however this agreement was marginally not statistically significantly greater than chance (k = 0.27, p = 0.07). PFA-100 CEPI CT and AA-induced aggregometry were concordant in 65% of the samples. This agreement was not statistically significant. When we determined the responses as normal, or pathological (1 cut-off point), the overall agreement between PFA-100 CEPI CTs and aggregometry with EPI was high (a = 81%, k = 0.59, p = 0.009). On the contrary, the agreement between PFA-100 CEPI CT and aggregometry with AA was very poor. Analysis of the data, using the non-parametric Spearman rank correlation coefficient, revealed that PFA-100 CEPI CT was inversely correlated to VWF Actity (rho = 0.55, p = 0.022). In similar, PFA-100 CADP CT was statistically significantly inversely correlated to PLT count (rho = 0.69, p = 0.002). Last, PLT count was inversely correlated to VWF Activity/ VWF: Ag (rho = 0.52, p = 0.031).

Discussion
Thrombotic complications are the major cause of morbidity and mortality in ET and the more frequent arterial thrombi are composed largely of platelets. At low platelet counts, aspirin provides satisfactory protection against thrombotic episodes and is therefore considered first line therapy [17]. However, at higher platelet counts, ASA has been less effective. A dose-effect phenomenon or an acquired resistance to aspirin have been suggested as potential underlying

Figure 2 Comparison of all samples (n = 33) tested by the PFA-100 CEPI cartridge and by optical aggregometry using arachidonic acid (AA) 500 g/ML as an agonist. PFA-100 CT (164 sec or aggregation b 20% (indicated by the lines on the graph) was determined as pathological platelet response.

Evaluation of PFA-100 in monitoring aspirin treatment in thrombocytosis mechanisms for this phenomenon [18]. On the other hand, ASA probably cause bleeding complications in patients with ETat a frequency higher than in healthy individuals [19]. Thus, it would be of great value to be able to monitor antiaggregating properties of aspirin with a valid and rapid method and subsequently establish the potential relevance of these measurements with the clinical outcome. Platelet aggregation has already been used successfully for the assessment of the haemostatic balance in chronic myeloproliferative disorders, including ET [13,20]. In parallel, it has been applied to define the response to aspirin in several groups of patients [14,21,22]. Although it is an old and probably the most widely used test to measure platelet function, it is time consuming, poorly standardized [23] and conflicting views have been expressed regarding its reproducibility and accuracy in monitoring aspirin effect [24,25]. On the other hand, the PFA-100 was found to be a sensitive, rapid and easy way to detect aspirin-induced platelet dysfunction [26,27], although a strong influence on CTs by VWF levels has been reported [28]. To our knowledge, only two studies have previously examined the role of PFA-100 in the evaluation of platelet function in ET patients. One study has compared the platelet function measured by classic aggregation tests with the closure time obtained by the PFA-100 in ET patients not receiving any antiplatelet drug [15], while the other study compared the influence of aspirin on the Ivy bleeding time and the PFA-100 closure time in five ET patients being on aspirin [29]. This is the first study to compare platelet aggregometry and PFA-100 as screening tools for aspirin responsiveness in patients with increased platelet counts. Our findings showed that PFA-100 CEPI CTand EPIinduced aggregometry is the pair of methods with the higher agreement in monitoring of platelet dysfunction due to ASA treatment. Satisfactory results were also obtained in patients not taking aspirin. The comparison between PFA-100 CEPI CTs and AA-induced aggregometry exhibited moderate agreement in the total number of patients and in the group treated with ASA. Generally higher agreements were achieved, when 1 cut-off point had been used for comparisons. PFA-100 CADP CTs and ADPinduced aggregometry were not concordant. This is probably due to the fact that several variables affect the two methods in a different way [30,31]. On the other hand, it is worth of noting the satisfactory agreement achieved between PFA-100 CEPI CTs and EPI-induced aggregometry, even when two cut-off points had been selected for comparisons. It seems that there is a concordance between the two methods, even in the assessment of the degree of platelet

105

dysfunction due to aspirin treatment. Moreover, it is very interesting that this combination of methods also showed agreement in patients not treated with aspirin. Several studies have compared PFA-100 with optical aggregometry in patients receiving ASA and showed poor agreement between the two methods [14,21,32]. ADP and AA had been used as agonists in optical aggregometry in these studies. Malinin et al. used aggregometry with EPI 5 M along with PFA-100 to monitor aspirin therapy and found a high correlation between the methods [22]. Thus, it is confirmed that if aggregometry is used to define the response to antiaggregating agents, agreement on the kind of inducer, its concentrations and the selection of cutoff points is essential [23]. Testing the correlation of the several variables known to influence PFA-100 function with closure times, we corroborated the significant effect (inverse correlation) of the VWF activity levels on PFA-100 measurements. It is noteworthy that regarding CEPI cartridges, this influence was observed in patients not treated with ASA and in the total group. On the contrary, in the case of CADP cartridges the same correlation was evident in patients treated with aspirin and in the total group of patients. However, it was marginally not statistically significant in the group of patients not taking ASA. This finding shows that even in higher platelet counts the CEPI cartridge remains more sensitive to antiaggregating properties of aspirin than CADP cartridge, which in turn, can better detect VWF defects when these coexist with platelet dysfunction due to aspirin intake. Another interesting point was that PLT count was inversely correlated to VWF Activity / VWF :Ag. This is in accordance with the view that acquired abnormalities of plasma VWF are related to the platelet count [33]. Last, a correlation between PLT counts and the PFA-100 measurements could not be established. However, CADP CT in patients not taking aspirin was statistically significantly inversely correlated to PLT. Our study has several limitations. The fact that the highest platelet count in our patients was 1,200,000/L due to hydroxyurea treatment given to most of patients did not allow us to study the PFA100 system performance characteristics in higher thrombocytosis levels. The causes for flow obstruction, which is a significant limitation of the method, in about 10% of samples tested with the PFA-100 system, were not elucidated. It could be also argued that we should have used controls with platelet counts less than 500,000/L, in order to increase the validity of our findings. However, several studies have compared the two methods in patients treated with ASA and having normal platelet counts, so we considered that our findings in such a group would

106 not add to the literature. Last, our study population is heterogeneous because both ETand RT patients were included. However, we believe that this contributes to the better assessment of the influence of thrombocytosis on the PFA-100 measurements, since platelets in ET present several functional abnormalities not related with antiplatelet therapy. In conclusion, it seems that the PFA-100 system is a reliable and rapid method in the assessment of antiplatelet effect of aspirin in patients with thrombocytosis, especially in association with EPI or AAinduced aggregometry. The clinical usefulness of these measurements with the PFA-100 system and optical aggregometry has not been estimated and such an attempt was beyond the aims of this study. However, it would be of great interest to determine whether there is any relevance between the severity of platelet dysfunction due to aspirin intake as monitored by these methods and clinical outcome in thrombocythemic patients.

A.E. Tsantes et al.

[9] Escolar G, Cases A, Vinas M, Pino M, Calls J, Cirera I, et al. Evaluation of acquired platelet dysfunctions in uremic and cirrhotic patients using the platelet function analyzer (PFA100): influence of hematocrit elevation. Hematologica 1999;84:6149. [10] Homoncik M, Jilma B, Hergovich N, Stohlawetz P , Panzer S, Speiser W. Monitoring of aspirin (ASA) pharmacodynamics with the platelet function analyzer PFA-100. Thromb Haemost 2000;83:31621. [11] Harrison P. The role of PFA-100 testing in the investigation and management of haemostatic defects in children and adults. Br J Haematol 2005;130:310. [12] Murphy S, Peterson P, Iland H. Experience of the Polycythemia Vera Study Group with essential thrombocythemia: a final report on diagnostic criteria, survival, and leukemic transition by treatment. Semin Hematol 1997;34: 2939. [13] Avram S, Lupu A, Angelescu S, Olteanu N, Mut-Popescu D. Abnormalities of platelet aggregation in chronic myeloproliferative disorders. J Cell Mol Med 2001;5:7987. [14] Gum PA, Kottke-Marchant K, Poggio ED, Gurm H, Welsh PA, Brooks LM, et al. Profile and prevalence of aspirin resistance in patients with cardiovascular disease. Am J Cardiol 2001;88: 2305. [15] Cesar JM, de Miguel D, Avello AG, Burgaleta C. Platelet dysfunction in primary thrombocythemia using the Platelet Function Analyzer, PFA-100. Am J Clin Pathol 2005;123:7727. [16] Sucker C, Senft B, Scharf R, Zotz R. Determination of von Willebrand factor activity: Evaluation of the Haemosil assay in comparison with established procedures. Clin Appl Thromb/Hemost 2006;12:30510. [17] van Genderen PJ, Mulder PG, Waldeboer M. Prevention and treatment of thrombotic complications in ET: efficacy and safety of aspirin. Br J Haematol 1997;97:17984. [18] Petrides EP, Siegel F. Thrombotic complications in essential thrombocythemia (ET): Clinical facts and biochemical riddles. Blood Cells Mol Dis 2006;36:37984. [19] Cortelazzo S, Marchetti M, Orlando M. Aspirin increases the bleeding side effects in essential thrombocythemia independent of the cyclooxygenase pathway: role of the lipooxygenase pathway. Am J Hematol 1998;57:27782. [20] Raman BK, Van Slyck E, Riddle J, Sawdyk MA, Abraham JP, Saeed S. Platelet function and structure in myeloproliferative disease, myelodysplastic syndrome, and secondary thrombocytosis. Am J Clin Pathol 1989;91:64755. [21] Harrison P, Segal H, Blasbery K, Furtado C, Silver L, Rothwell PM. Screening for aspirin responsiveness after transient ischemic attack and stroke. Comparison of 2 point-of-care platelet function tests with optical aggregometry. Stroke 2005;36:10015. [22] Malinin AI, Atar D, Callahan K, McKenzie M, Serebruany VL. Effect of a single dose of aspirin on platelets in humans with multiple risk factors for coronary artery disease. Eur J Pharmacol 2003;462:13943. [23] Breddin HK. Can platelet aggregometry be standardized? Platelets 2005;16:1518. [24] De Gaetano G, Cerleti C. Aspirin resistance: a revival of platelet aggregation tests? J Thromb Haemost 2003;1: 204850. [25] Patrono C. Aspirin resistance: definition, mechanisms and clinical reads-outs. J Thromb Haemost 2003;1:17103. [26] Francis J, Francis D, Larson L, Helms E, Garcia M. Can the Platelet Function Analyzer (PFA)100 test substitute for the template bleeding time in routine clinical practice? Platelets 1999;10:1326. [27] Marshall PW, Williams AJ, Dixon RM, Growcott JW, Warburton S, Armstrong J, et al. A comparison of the effects of

Acknowledgments
G. Mantzios was a gifted and outstanding physician that recently passed away at the age of 39. We are grateful to him not only for his significant contribution to the attainment of this study, but, mainly, for serving as a role model to us in our professional and personal lives. He will be sorely missed. Financial support: This work was supported by University of Athens (grant 70/4/9107 to A. Tsantes).

References
[1] Bellucci S, Janvier M, Tobelem G. Essential thrombocythemia: clinical evolutionary and biologic data. Cancer 1986;58: 24407. [2] Harrison CN. Platelets and thrombosis in myeloproliferative diseases. Hematol Am Soc Hematol Educ Program 2005: 40915. [3] Elliott MA, Tefferi A. Thrombosis and haemorrhage in polycythaemia vera and essential thrombocythaemia. Br J Haematol 2005;128:27590. [4] van Genderen PJ, van Vliet H, Prins FJ, van de Moesdijk D, van Strick R, Zijlstra FJ, et al. Excessive prolongation of the bleeding time by aspirin in essential thrombocythemia is related to a decreased of large von Willebrand factor multimers in plasma. Am Hematol 1997;75:21520. [5] Kundu SK, Heilmann EJ, Sio R, Garcia C, Davidson RM, Ostgaard RA. Description of an in vitro platelet function analyzer PFA-100. Semin Thromb Hemost 1995;21:10612. [6] Jilma B. Platelet function analyzer PFA-100: a tool to quantify congenital or acquired platelet dysfunction. J Lab Clin Med 2001;138:15263. [7] Cattaneo M, Federici AB, Lecchi A. Evaluation of PFA-100 system in the diagnosis and therapeutic monitoring of patients with von Willebrand disease. Thromb Haemost 1999;82: 359. [8] Cattaneo M, Lecchi A, Agati B. Evaluation of platelet function with the PFA-100 system in patients with congenital defects of platelet secretion. Thromb Res 1999;96: 2137.

Evaluation of PFA-100 in monitoring aspirin treatment in thrombocytosis

aspirin on bleeding time measured using the PFA-100, in healthy volunteers. Br J Clin Pharmacol 1997;44:1515. [28] Chakroun T, Gerotziafas G, Robert F, Lecrubier C, Samama MM, Hatmi M, et al. In vitro aspirin resistance detected by PSA-100TM closure time: pivotal role of plasma von Willebrand factor. Br J Haematol 2004;124:805. [29] Troost MM, van Genderen PJ. Excessive prolongation of the Ivy bleeding time after aspirin in essential thrombocythemia is also demonstrable in vitro in the high shear stress system PFA-100. Ann Hematol 2002;81:3523. [30] Favaloro EJ. Utility of the PFA-100 for assessing bleeding disorders and monitoring therapy: a review of analytical

107

variables, benefits and limitations. Haemophilia 2001;7: 1709. [31] Storey RF, Heptinstall S. Laboratory investigation of platelet function. Clin Lab Haematol 1999;21:31729. [32] McDabe DJ, Harrison P, Mackie IJ, Sidhu PS, Lawrie AS, Purdy G, et al. Assessment of the antiplatelet effects of low to medium dose aspirin in the early and late phases after ischaemic stroke and TIA. Platelets 2005;16:26980. [33] van Genderen PJ, Budde U, Michiels J, van Strik R, van Vliet H. The reduction of large von Willebrand factor multimers in plasma in essential thrombocythemia is related to the platelet count. Br J Haematol 1996;93:9625.