Biological effects of aspirin and clopidogrel in a randomized cross-over study in 96 healthy volunteers

Authors

  • P. FONTANA,

    1. Department of Internal Medicine, Faculty of Medicine, Division of Angiology and Hemostasis, University Hospitals of Geneva, Geneva, Switzerland
    Search for more papers by this author
  • S. NOLLI,

    1. Department of Internal Medicine, Faculty of Medicine, Division of Angiology and Hemostasis, University Hospitals of Geneva, Geneva, Switzerland
    Search for more papers by this author
  • G. REBER,

    1. Department of Internal Medicine, Faculty of Medicine, Division of Angiology and Hemostasis, University Hospitals of Geneva, Geneva, Switzerland
    Search for more papers by this author
  • P. DE MOERLOOSE

    1. Department of Internal Medicine, Faculty of Medicine, Division of Angiology and Hemostasis, University Hospitals of Geneva, Geneva, Switzerland
    Search for more papers by this author

Pierre Fontana, Division of Angiology and Hemostasis, University Medical Center, 1 rue Michel-Servet, CH-1211 Geneva 14, Switzerland.
Tel.: +41 22 379 55 67; fax: +41 22 372 92 99; e-mail: pierre.fontana@medecine.unige.ch

Abstract

Summary. Background: Some data suggest that biological ‘resistance’ to aspirin or clopidogrel may influence clinical outcome. Objective: The aim of this study was to evaluate the relationship between aspirin and clopidogrel responsiveness in healthy subjects. Methods: Ninety-six healthy subjects were randomly assigned to receive a 1-week course of aspirin 100 mg day−1 followed by a 1-week course of clopidogrel (300 mg on day 1, then 75 mg day−1), or the reverse sequence, separated by a 2-week wash-out period. The drug effects were assessed by means of serum TxB2 assay, platelet aggregation tests, and the PFA -100® and Ultegra RPFA -Verify Now® methods. Results: Only one subject had true aspirin resistance, defined as a serum TxB2 level > 80 pg μL−1 at the end of aspirin administration and confirmed by platelet incubation with aspirin. PFA-100® values were normal in 29% of the subjects after aspirin intake, despite a drastic reduction in TxB2 production; these subjects were considered to have aspirin pseudo-resistance. Clopidogrel responsiveness was not related to aspirin pseudo-resistance. Selected polymorphisms of platelet receptor genes were not associated with either aspirin or clopidogrel responsiveness. Conclusions: In healthy subjects, true aspirin resistance is rare and aspirin pseudo-resistance is not related to clopidogrel responsiveness.

Introduction

Inhibiting platelet function with aspirin or clopidogrel is an effective way of preventing cardiovascular events [1,2]. Aspirin and clopidogrel target two distinct platelet amplification pathways necessary for full platelet activation. Aspirin specifically inhibits thromboxane (Tx) A2 production by irreversibly acetylating a serine residue at position 529 of the cyclooxygenase-1 (COX-1) protein [3], while clopidogrel, a thienopyridine, targets the adenosine-diphosphate (ADP) pathway by irreversibly blocking the ADP receptor P2Y12 [4].

Despite appropriate antiplatelet therapy, vascular events recur in a significant proportion of patients, raising the possibility of biological ‘antiplatelet drug resistance’ being implicated in these treatment failures [5]. There is no consensus definition of aspirin resistance, and its incidence ranges from 5% when based on a specific test (TxA2 production assay) to 56% when based on a non-specific test (e.g. PFA-100®) [6]. Aspirin resistance as defined with either specific or non-specific tests has been associated with clinical events [7–12]. Clopidogrel responsiveness is usually based on the inhibition of ADP-induced platelet aggregation [5], and biological poor responsiveness to this drug has been associated with recurrence of cardiovascular events in one recent study [13].

The mechanisms underlying ‘antiplatelet drug resistance’ are not known, but may include genetic variability. Among the genetic polymorphisms studied, the PlA1/PlA2 of glycoprotein (GP) IIIa has been suggested to be associated with both aspirin and clopidogrel poor responsiveness [14,15]. Thus, one could speculate that aspirin and clopidogrel biological effects may be linked, and as clopidogrel has been proposed as an alternative for aspirin-resistant patients [16], this issue is of crucial importance.

Methods

Subjects

Ninety-six unrelated healthy Caucasian men aged from 18 to 40 years were recruited. They were all non-smokers and denied taking any medication for at least 10 days before the start of the study. In an open-labeled randomized design, the subjects received a 1-week course of aspirin 100 mg day−1 followed (n = 45) or preceded (n = 51) by a 1-week course of clopidogrel (300 mg on the first day, then 75 mg day−1). The two courses were separated by a 2-week wash-out period. Adherence was evaluated by pill counting and face-to-face interview. Blood was sampled at baseline and on the morning of the day after the end of each 7-day treatment course. At baseline, platelet function tests and routine laboratory tests comprised platelet aggregation, the PFA-100® closure time (CT), serum assay of TxB2 (the stable breakdown product of TxA2). von Willebrand ristocetin cofactor assay (VWF) was performed by a automated turbidimetric method (Dade Behring, Marburg, Germany). After each treatment period, the platelet function tests depended on the drug administered during the previous 7 days, that is, serum TxB2, platelet aggregation, PFA-100® CT, and the Ultegra RPFA -Verify Now® ASA (RPFA, Accumetrics Inc., San Diego, CA, USA) after the aspirin course, and platelet aggregation and serum TxB2 after the clopidogrel course. Two subjects were lost to follow up after receiving aspirin but before receiving clopidogrel. Thus, the aspirin effect was evaluated in 96 subjects and the clopidogrel effect in 94 subjects.

All subjects gave written informed consent, and the study protocol was approved by the ethics committee of University Hospitals of Geneva.

Sample collection and platelet function tests

Venous blood was collected using a 19-gauge needle and no tourniquet after an overnight fast, in tubes containing either no anticoagulant or 0.105 m sodium citrate (1 vol./9 vol.) (BD Vacutainer®; Becton Dickinson, Meylan, France). An additional blood sample was taken in a special citrated tube for RPFA analysis (Accumetrics Inc). The first 3 mL of blood was discarded.

Platelet aggregation was evaluated in citrated platelet-rich plasma (PRP) on a four-channel aggregometer (Bio/Data PAP-4; Bio/Data Corporation, Horsham, PA, USA), using one of the following agonists: arachidonic acid (AA) 1.5 mm (Bio/Data Corporation), Horm collagen 0.5 μg mL−1 (Nycomed, Munich, Germany) or ADP 20 μm (Helena Biosciences Europe, Sunderland, UK).

TxA2 production in response to endogenous thrombin was evaluated by allowing a 6 mL tube of whole blood to clot at 37 °C for 2 h, as previously described [17]. Serum was stored at −80 °C until use, within 3 months after collection, by using a commercial ELISA kit (Amersham Biosciences, Otelfingen, Switzerland) blindly to the other test results.

PFA-100® and the Ultegra RPFA-Verify Now®ASA (RPFA). The PFA-100® device (Dade-Behring, Düdingen, Switzerland) evaluates platelet function in high shear stress conditions in citrated whole blood. The CT was evaluated up to 300 s, the time limit of the device using the collagen/epinephrin cartridge.

RPFA (Accumetrics Inc.) is a rapid platelet-function assay based on cartridges containing propyl gallate as agonist. Results are expressed in aspirin response units (ARU). The cut-off of the RPFA method is 550 ARU, and aspirin non-responders would thus be expected to have values exceeding 550 ARU [18]. RPFA values were not available for the first 32 subjects.

Evaluation of drug responses

Aspirin  Measurement of TxA2 production in terms of the serum TxB2 level was used as the gold standard for the biological effect of aspirin. The distribution of baseline serum TxB2 concentrations was used to determine a cut-off value. True aspirin resistance was defined, after the 1-week course of aspirin, by a serum TxB2 concentration above the 5th percentile of the baseline values. The non-aspirin-specific test PFA-100® was used to define non-specific aspirin resistance. A cut-off was derived from the distribution of baseline values. Non-specific aspirin resistance was defined by a PFA-100® value below the 95th percentile of the baseline values after aspirin administration.

Clopidogrel  ADP-induced maximal platelet aggregation (referred to below as ADP aggregation) was used to evaluate the effect of clopidogrel. A concentration of 20 μm was preferred, as this concentration has been used in several studies of clopidogrel responsiveness and that 20μm ADP aggregation is not affected by the TxA2-dependant amplification of platelet aggregation when intermediate concentrations are used in normal citrated PRP [5]. Subjects were stratified into four quartiles based on the percentage reduction in ADP aggregation from baseline after the 1 week course of clopidogrel [13].

Aspirin dose–effect relationship and in vitro effects

In six selected individuals, the effect of a single daily aspirin dose of 100, 300, or 500 mg for three consecutive days was evaluated on the basis of serum TxB2, AA-induced maximal platelet aggregation (referred to below as AA aggregation) and RPFA, between 2 and 10 months after the cross-over study. Aspirin intake was witnessed by an investigator and biological tests were carried out both before aspirin ingestion (day 0) and 24h after each dose (days 1, 2, and 3).

In the same individuals, in vitro aspirin responses were evaluated in terms of AA aggregation on Day 0. PRP was incubated with water or aqueous solutions of 50, 100 and 200 μm aspirin (dl-lysine-acetylsalicylate; Sanofi-Synthelabo, Geneva, Switzerland) at 37 °C for 2 min before adding 1.5 mm AA.

Flow cytometry

In these six individuals, platelet COX-1 and -2 expression was evaluated by means of flow cytometry as described elsewhere [19]. Permeabilized washed platelets were incubated with fluorescein isothiocyanate-conjugated monoclonal anti-COX-1 (Cayman Chemical, Ann Arbor, MI, USA), or monoclonal anti-COX-2 (Cayman Chemical), or a negative isotype control (Becton Dickinson). Platelets were identified by positive staining with PE -conjugated anti-CD42b antibodies (Dakocytomation, Baar, Switzerland). Ten thousand events were acquired on a FACStrack flow cytometer (Becton Dickinson), and data were analyzed using cellquest software (Becton Dickinson).

Genotyping studies

P2Y12 (H1/H2), GPIIIa (PlA1/PlA2), and GPIa (C807T) gene polymorphisms were identified by restriction fragment length polymorphism analysis, as described elsewhere [20–22].

Genotyping results were analyzed blindly to phenotypic status.

Statistical analysis

Data are shown as means ± SD unless otherwise stated. Categorical variables are expressed as percentages. The chi-squared test was used to test the association between drug response categories and selected genetic polymorphisms. Trends of continuous variables across genotype and drug response groups were tested with the non-parametric Kruskall–Wallis test. Multivariate logistic regression models were used to test the relationship between selected log-transformed variables and the aspirin and clopidogrel responses.

Statistical tests were performed with the stata® 7.0 software package (Stata Corp., College Station, TX, USA), and differences with P-values < 0.05 were considered significant.

Results

Aspirin effect

The median serum TxB2 level at baseline was 322 pg μL−1 (interquartile range [IR]: 209–402 pg μL−1). Aspirin had a drastic effect on TxB2 levels, with a median value of 3 pg μL−1 (IR: 2–5 pg μL−1). The 5th percentile of baseline TxB2 levels, used to define true aspirin resistance in this study, was 80 pg μL−1. Only one subject [1%, 95% confidence interval (CI): 0–6%] had true aspirin resistance. The TxB2 values in this subject were 360 pg μL−1 at baseline and 113 pg μL−1 after a 1-week course of aspirin 100 mg day−1. Similar results were obtained if TxB2 levels were adjusted to platelet count (data not shown).

Arachidonic acid aggregation fell from a median of 70% at baseline (IR: 61% and 79%) to 4% (IR: 2% and 7%) after a 1-week course of aspirin. The subject with true aspirin resistance had AA aggregation values of 68% at baseline and 76% after aspirin administration, and showed an irreversible aggregation profile on both occasions. The other 95 subjects had AA aggregation values ranging from 0% to 17% after aspirin administration.

The RPFA values after aspirin administration were 581 ARU in the subject with true aspirin resistance and 383 ± 46 ARU in the other subjects. The subject with true aspirin resistance was the only subject with an RPFA value exceeding 550 ARU.

The baseline PFA-100® CT was 128 ± 31 s. The 95th percentile of baseline PFA-100® CT values, used as the cut-off for non-specific aspirin resistance, was 190 s. After aspirin exposure, the PFA-100® CT ranged from 103 to > 300 s, and 29 subjects (30.2%) had values below 190 s. The subject with true aspirin resistance had PFA-100® CT values of 109 and 103 s before and after aspirin exposure, respectively. Subjects with PFA-100® CT < 190 s and serum TxB2 < 80 pg μL−1 after aspirin exposure [n = 28, 29%, CI (21–40%)] were defined as ‘aspirin pseudo-resistant’ (APR). The other 67 subjects, who had PFA-100® CT values above 190 s after the aspirin course, had serum TxB2 concentrations below 80 pg μL−1 and were defined as aspirin-sensitive (AS). The characteristics of APR and AS subjects are described in Table 1. Baseline PFA-100® CT, VWF, and collagen lag time values differed significantly between APR and AS subjects. Multivariate logistic regression analysis with the platelet count, hematocrit, fibrinogen, VWF, basal PFA-100® CT, and basal collagen lag time as independent variables showed that VWF and the collagen lag time were the only variables independently associated with APR and AS status with respective odds ratios (OR) of 1.02 (CI: 1.01–1.03, P = 0.003) and 0.95 (CI: 0.91–0.99, P = 0.016).

Table 1.  Baseline characteristics of aspirin pseudo-resistant (APR) and aspirin-sensitive (AS) subjects
 APR n = 28AS n = 67P
  1. BMI, body mass index; PFA, platelet function analyzer; CT, closure time; TxB2, thromboxane B2; AA, arachidonic acid; ADP, adenosine diphosphate; VWF, von Willebrand factor.

Age (years)29.3 ± 5.827.5 ± 5.40.2
BMI (kg m−2)23.5 ± 3.023.2 ± 2.40.5
PFA -100® CT (s)116.6 ± 19.7133.2 ± 34.10.02
TxB2 (pg μL−1)338 ± 160326 ± 1750.7
AA aggregation (%)69.3 ± 15.168.0 ± 17.80.9
Collagen lag time (s)50.2 ± 10.260.4 ± 19.70.02
ADP aggregation (%)74.5 ± 8.277.2 ± 13.30.6
VWF (%)97.5 ± 37.570.7 ± 35.8< 0.01
Fibrinogen (g L−1)2.4 ± 0.42.4 ± 0.50.9
Hematocrit (%)42.5 ± 1.942.8 ± 1.90.7
Platelet count (G L−1)214.9 ± 31.8213.9 ± 43.80.9

The influence of the aspirin dose was evaluated, some time after the cross-over study, in the subject with true aspirin resistance and in five randomly selected subjects (controls) who had serum TxB2 values below 80 pg μL−1 after the 1-week aspirin course. After a single dose of 100 mg aspirin, the mean serum TxB2 fell below 80 pg μL−1 in the control group (53 ± 24 pg μL−1) and remained high (115 pg μL−1) in the subject with true aspirin resistance ( Fig. 1A, Day 1). Similarly, AA aggregation was 44% ± 31% in the controls and 85% in the subject with true aspirin resistance (Fig. 1B, Day 1). In contrast, after aspirin doses of 300 mg and 500 mg, TxB2 production and AA aggregation fell markedly in all six subjects (Fig. 1A,B, days 2 and 3). Although TxB2 values on days 2 and 3 of treatment with these doses remained low in the subject with true aspirin resistance, this subject's RPFA values were 516 and 586 ARU, respectively (Fig. 1C).

Figure 1.

Effect of an aspirin dose increment in the subject with true aspirin resistance (open squares) and in five randomly selected controls (closed squares), based on serum TxB2 assay (A), arachidonic-induced maximal platelet aggregation (AA aggregation, B), and RPFA assay (C). Results on days 1, 2 and 3 were obtained 24 h after 100, 300, and 500 mg aspirin intake, respectively.

The effect of aspirin was tested in vitro on day 0, before aspirin intake, in the same six subjects. After incubation with 100 μm aspirin, AA aggregation was still 70% in the subject with true aspirin resistance, compared with 6% ± 1% in the controls ( Fig. 2). As in the ex vivo study, AA aggregation in the subject with true aspirin resistance fell to a value similar to that in the controls when the aspirin concentration was increased.

Figure 2.

Arachidonic-induced maximal platelet aggregation (AA aggregation) after incubation of platelet-rich plasma with various concentrations of aspirin, in the subject with true aspirin resistance (open square) and in five randomly selected controls (closed squares).

COX quantification

The subject with true aspirin resistance had platelet COX-1 and -2 mean fluorescence values of 27.7 and 63.3, respectively, compared with the control group with values of 27.3 ± 6.5 and 79.2 ± 23.6 for COX-1 and -2 expression, respectively.

Clopidogrel effect

After a 1-week clopidogrel course, the clopidogrel response was stratified into four quartiles. Subjects in the first and second quartiles had ADP aggregation values of 65% ± 13% and 45% ± 3%, respectively, while subjects in the 3rd and 4th quartiles had values of 36% ± 3% and 23%±5% (P < 0.001). Serum TxB2 concentration was not affected by clopidogrel intake with median value of 309 pg μL−1 (IR: 209 to 381 pg μL−1, P = NS compared with basal TxB2 level). Table 2 shows the characteristics of the subjects according to the quartile of clopidogrel responsiveness. The baseline collagen lag time increased significantly with clopidogrel effect. Multivariate logistic regression analysis, in which quartiles 1 and 2 were coded 1 and quartiles 3 and 4 were coded 0, showed that this association (OR 0.96, CI: 0.93–0.99, P = 0.016) remained significant after adjustment for the platelet count, the hematocrit, and the fibrinogen and VWF levels. Interestingly, although the baseline ADP aggregation was not associated with the clopidogrel effect, ADP aggregation evaluated after the clopidogrel course correlated negatively with the clopidogrel effect: values were 47.1% ± 12.6%, 34.3% ± 6.0%, 27.5% ± 5.5% and 17.2% ± 4.2% in quartiles 1 through 4, respectively (P < 0.001). When added to the multivariate model, postclopidogrel ADP aggregation was the only variable significantly associated with the clopidogrel effect, with an OR of 1.3 (CI: 1.2–1.5, P < 0.001).

Table 2.  Baseline characteristics of the subjects according to clopidogrel responsiveness
 Quartile of clopidogrel responsivenessP
1 (n = 23)2 (n = 23)3 (n = 24)4 (n = 24)
  1. BMI, body mass index; PFA, platelet function analyzer; CT, closure time; TxB2, thromboxane B2; AA, arachidonic acid; ADP, adenosine diphosphate; VWF, von Willebrand factor.

Age (years)29.3 ±  5.327.6 ±  4.826.9 ± 6.428.2 ± 5.50.5
BMI (kg m−2)23.0 ± 1.623.3 ± 2.723.7 ± 2.623.2 ± 3.20.6
ADP aggregation (%)71.7 ± 10.175.2 ± 10.576.7 ± 13.474.0 ± 13.00.4
Collagen lag time (s)50.7 ± 13.052.7 ± 9.955.7 ± 13.168.8 ± 26.50.03
AA aggregation (%)67.0 ± 16.872.2 ± 10.469.8 ± 16.163.9 ± 21.60.4
VWF (%)82.9 ± 37.281.2 ± 44.576.0 ± 29.781.0 ± 43.40.9
Fibrinogen (g L−1)2.4 ± 0.52.5 ± 0.52.4 ± 0.62.4 ± 0.30.8
Hematocrit (%)42.3 ± 1.942.6 ± 2.242.6 ± 1.943.2 ± 1.40.3
Platelet count (G L−1)204 ± 47216 ± 44217 ± 30215 ± 430.8

Aspirin effect is expressed according to the quartile of clopidogrel effect in Table 3. There was no association between the clopidogrel effect and the TxB2 level or the proportion of subjects with aspirin pseudo-resistance. The subject with true aspirin resistance was in quartile 1 of clopidogrel responsiveness, with a postclopidogrel ADP aggregation value of 101% of baseline.

Table 3.  Aspirin responsiveness according to the quartile of clopidogrel responsiveness
 Quartile of clopidogrel responsivenessP
1 (n = 23)2 (n = 23)3 (n = 24)4 (n = 24)
  1. RPFA, Ultegra RPFA-Verify Now®; IR, interquartile range.

TxB2 (pg μL−1)8.4 ± 23.04.1 ± 2.04.4 ± 2.56.3 ± 12.90.4
AA (%)8.6 ± 15.05.6 ± 4.14.4 ± 4.24.1 ± 4.40.08
PFA-100® (s, median and IR )220 (155 to > 300)300 (155 to > 300)300 (167 to > 300)295 (161to > 300)0.6
RPFA (ARU)393 ± 65388 ± 44391 ± 60369 ± 400.6
APR subjects (n, %)7 (29)7 (29)6 (27)8 (36)0.8

Genotyping results

Table 4 shows clopidogrel and aspirin potency according to common genetic variations suggested to be associated with aspirin or clopidogrel responsiveness. There was no significant association between these genetic variations and the potency of either drug. The subject with true aspirin resistance was T/T, PLA2/A2, and H1/H1 for the C807T (GPIa) PLA1/A2 (GPIIIa), and H1/H2 (P2Y12) polymorphisms, respectively.

Table 4.  Effects of aspirin and clopidogrel and distribution of APR subjects according to selected genetic polymorphisms. The genetic profile of the subject with true aspirin resistance is described in the text
Gene (SNP )GenotypenTxB2 post aspirin (pg μL−1)PAPR subjects n (%)PClopidogrel effect (% of baseline)P
  1. SNP, single nucleotide polymorphism.

GPIa (C807T)C/C344 ± 30.248 (24)0.3442.3 ± 14.00.84
C/T456 ± 913 (29)40.6 ± 16.6
T/T163 ± 27 (44)42.9 ± 18.4
GPIIIa (PlA1/PLA2)PLA1/A1654 ± 30.3118 (28)0.8043.0 ± 15.30.32
PLA1/A2256 ± 128 (32)39.7 ± 18.0
PLA2/A253 ± 12 (40)34.1 ± 10.2
P2Y12 (H1/H2)H1/H1705 ± 80.8322 (31)0.4841.3 ± 15.40.95
H1/H2254 ± 36 (24)42.5 ± 17.6

Discussion

This study shows that true aspirin resistance is rare in healthy subjects, as it has been reported in patients [23]. Aspirin pseudo-resistance, defined by a PFA-100® CT value of < 190 s and a serum TxB2 value of < 80 pg μL−1, is more frequent. The rate of aspirin pseudo-resistance was 29% (CI: 21–40%), a proportion similar to that found in studies of patients, in whom aspirin non-responsiveness was based on the PFA-100® CT value only [24, 25]. Aspirin is equally potent in APR subjects and in aspirin sensitive subjects (TxB2 production is equally abolished by aspirin in both groups), a phenomenon already described [26]. However, in APR platelets, the inhibition of the TxA2 pathway by aspirin may not be sufficient to adequately inhibit global platelet function. Indeed, platelets of subjects with aspirin pseudo-resistance were more sensitive to collagen, as previously reported in a study of eight healthy volunteers [27]. As collagen-induced platelet aggregation is dependent on several amplification pathways [28], the inhibition of the TxA2 amplification loop could be compensated for by over-expression of other amplification pathways that are insensitive to aspirin, such as CD40L [29], in addition to high VWF. As several studies linked either high VWF or low PFA-100® CT values with vascular events in aspirin-treated patients [9–12, 30], aspirin pseudo-resistance should be considered as a potential risk factor for recurrence of cardiovascular events.

The data concerning true aspirin resistance have to be evaluated with caution as we found only one subject in our population. However, this subject had a serum TxB2 value of 113 pg μL−1 and 76% AA aggregation after 1 week course of aspirin, a phenotype consistent with the in vitro study, suggesting a persistent COX activity despite aspirin intake. Platelet COX-2 has been implicated in aspirin non-responsiveness [19]. However, we found that platelet COX-2 expression did not differ substantially between the resistant subject and controls. In order to rule out a mutation of COX-1 serine 529, this region was sequenced in the resistant subject and in the five control subjects. No sequence variations were found in the resistant subject (data not shown). Six months later, a single dose of 100 mg aspirin produced a similar effect (TxB2 115 pg μL−1). In the five controls tested concomitantly, the mean serum TxB2 value was 53 ± 24 pg μL−1, compared with only 4 ± 2 pg μL−1 after the previous 1 week course of aspirin. Finally, increasing the aspirin dose, in vitro or ex vivo, abolished TxB2 production in the resistant subject. This points to a difference in pharmacokinetics between the resistant subject and the controls, possibly involving a plasma factor that largely neutralizes the biological effects of aspirin up to a given concentration threshold. Surprisingly, the RPFA value in the resistant subject was still > 550 ARU after a cumulative dose of 900 mg aspirin over 3 days, despite a low serum TxB2 level and a low AA aggregation, suggesting that the RPFA test might be sensitive to a factor involved in the mechanism of true aspirin resistance.

The effect of clopidogrel on platelet aggregation differed by a factor 3 between quartile 1 and quartile 4. Similar variability has been described in patients with ischemic heart disease [13]. The use of ADP aggregation to evaluate clopidogrel responsiveness has been criticized, and it has been suggested that measurement of the phosphorylation of vasodilator-stimulated phosphoprotein (VASP) may be more specific [5]. However, ADP aggregation correlates well with VASP phosphorylation [31]. Although subjects with aspirin pseudo-resistance and clopidogrel poor responsiveness had a short lag time of collagen-induced platelet aggregation, there was no relationship between clopidogrel responsiveness and aspirin pseudo-resistance, suggesting the involvement of different mechanisms. It is conceivable that VWF and aspirin-insensitive amplification pathways are involved in the short collagen lag time in subjects with aspirin pseudo-resistance, while high ADP sensitivity could contribute to the short collagen lag time in poor clopidogrel responders.

The subject with true aspirin resistance also had the weakest response to clopidogrel. Non-adherence to clopidogrel intake is a possibility, together with a chance association between true aspirin resistance and clopidogrel non-responsiveness in the same subject.

We found no association between the genetic polymorphisms listed in Table 4 and antiplatelet drug responsiveness in our healthy subjects. We chose to consider gene polymorphisms that have been implicated in aspirin or clopidogrel responsiveness with an allelic frequency that would allow us to detect an OR of approximately three with aspirin pseudo-resistance and a difference of 10% regarding clopidogrel responsiveness between carriers of at least one mutated allele and non-carriers (α = 0.05 and β = 0.2). Although a smaller effect of these polymorphisms cannot be ruled out, it is unlikely that the large interindividual variability in clopidogrel responsiveness in this study can be explained by the genetic variations listed in Table 4. However, patients, as opposed to healthy subjects, may have other factors acting in trans and modulating the effect of genetic variability on antiplatelet drug response. Thus, further studies should be performed with patients to draw definitive conclusions on the implication of genetic variability in antiplatelet drug response. A polymorphism of the COX-1 gene (C50T) was recently associated with aspirin responsiveness in patients [32] and in healthy subjects [33]. Considering the low allelic frequency of this polymorphism in Caucasian population, our study was underpowered to detect a significant difference according to the above-mentioned criteria.

The main limitation of this study was that the results apply only to healthy volunteers. However, genetic association studies with healthy subjects avoid the potential influence of confounding factors such as drug interactions or cardiovascular risk factors and provide interesting mechanistic clues.

In this study, we found that true aspirin resistance was rare among healthy subjects, and that it could be overcome by increasing the aspirin dose. Aspirin pseudo-resistance, defined in terms of PFA-100® CT and serum TXB2 values, is more frequent and may reflect a persistent ability of platelets to aggregate despite adequate inhibition of TxA2 production by aspirin, owing to high VWF levels and to compensatory platelet amplification pathways compensating for TxA2. Finally, aspirin pseudo-resistance was not related to clopidogrel responsiveness in our study. If aspirin pseudo-resistance and clopidogrel poor responsiveness are confirmed to be independent predictors of recurrence of cardiovascular events in patients, the present data should prompt a trial of such an antiplatelet drug switch during cardiovascular prophylaxis.

Acknowledgements

The authors thank Henri Bounameaux MD for his constant support, Pascale Gaussem PhD and Jean-Luc Reny MD, PhD for their helpful comments on the manuscript. Accumetrics provided the Ultegra-RPFA instrument. This work was supported by grants from the Internal Medicine Department of Geneva hospitals (PRD-03-II-18), the Swiss Society of Angiology, the Leenaards Foundation, and the Ingeborg Naegeli Foundation.

Ancillary