No Effect of Clopidogrel Activity or Cessation on Vascular Function or Markers of Inflammation

Abstract

The platelet adenosine diphosphate (ADP)-receptor blocker clopidogrel is effective in reducing the rate of thrombosis in cardiovascular disease, but may also have nonplatelet activity. However, there is variability in the suppression of platelet function in individuals, leading to the concept of clopidogrel resistance, that is, reduced platelet-suppressing activity. We tested the hypothesis that some of the beneficial effect of clopidogrel may be due to the variable activity of this drug on the vascular system (assessed by plasma markers von Willebrand factor and soluble E-selectin, and functional arterial pulse wave velocity) and inflammation (C-reactive protein and interleukin-6) while 32 patients with coronary artery disease taking 75 mg clopidogrel daily, and again 2 weeks after cessation of clopidogrel therapy. Platelet responsiveness to clopidogrel was assessed by the phosphorylation of intracellular regulatory protein—vasodilator-stimulated phosphoprotein method and aggregometry to ADP. Response to aspirin was assessed using arachidonic acid (AA), and soluble P-selectin and PAC-1 were also measured. While on clopidogrel, there were no relationships between any vascular or inflammatory index and the response to clopidogrel. After stopping clopidogrel, there were no differences in platelet aggregation to AA, or the expression of P-selectin or PAC-1 at rest, or after stimulation by AA, but platelet responses to ADP all increased (p < 0.01). Although soluble P-selectin increased when clopidogrel was stopped (p = 0.006), there were no changes in plasma markers or vascular function. We conclude that 75 mg/day clopidogrel has no effect of markers of vascular function or inflammation.

Clopidogrel is a key antiplatelet therapy in the reduction of the risk of thrombosis in coronary stenting and after myocardial infarction.^1,2,3 However, it has been suggested that the persistent occurrence of adverse clinical outcomes, including stent thrombosis after percutaneous coronary intervention, despite therapy with clopidogrel, may be due to a suboptimal response to this drug, known as high on-treatment platelet reactivity, or clopidogrel resistance.^4,5,6

Apart from its actions in suppressing platelet function, there are several reports suggesting that clopidogrel has nonplatelet effects that may contribute to its efficacy. These include mobilization of endothelial progenitor cells,⁷ improved vascular/endothelial function,^8,9,10 and reduced inflammation (as marked by high sensitivity [hs] C-reactive protein [CRP]).⁹ However, others reported evidence suggestive of an adverse effect of clopidogrel, such as a reduction in adenosine diphosphatase in vitro (which would favor platelet aggregation),¹¹ up-regulation of inflammatory gene expression,¹² and a decrease in CRP levels upon cessation.¹³ Few of these studies concurrently examined more than one marker in more than one particular potential pathophysiology, and none addressed the issue of a variable response to clopidogrel.

In an attempt to clarify these issues, we hypothesized (1) that selected markers of inflammation and vascular function would be influenced by differing reactivity toward clopidogrel, and (2) that the cessation of clopidogrel has a deleterious effect on those markers. We tested hypothesis (1) in a cohort of patients on long-term clopidogrel therapy, and hypothesis (2) after the same patients had stopped their clopidogrel.

Subjects and Methods

We recruited 32 patients (mean [standard deviation] age 60.7 [9.8] years, 25 men, 3 diabetic) with proven coronary artery disease potentially requiring the placement of at least one intracoronary stent, all of whom were prescribed 75 mg clopidogrel daily in addition to standard cardiovascular therapies. These therapies were aspirin by 31 patients, statin by 31, angiotensin converting enzyme inhibitor or angiotensin receptor blocker by 29, β blocker by 25, calcium channel blocker by 10, nitrate by 8, diuretic by 6, and insulin by 1. Cardiovascular histories were percutaneous coronary intervention in 31, myocardial infarction in 21, hypertension in 16, angina in 11, coronary artery bypass graft in 4, peripheral artery disease in 4, and stroke in 1. Duration of clopidogrel therapy was 1 year in 31 patients and 3 months in 1 patient. Exclusion criteria for all subjects were current use of oral or parenteral anticoagulation or other antiplatelet drugs, bleeding abnormalities, and/or significant hepatic, neoplastic, renal, or inflammatory disease. Patients were assessed while on clopidogrel, and again 2 weeks after they had stopped taking the drug in accordance with guidelines.³ No patient changed any other drug or its dose in the period while on and off clopidogrel. The project had the approval of the local research ethics committee and written informed consent was obtained from each subject.

Venous blood was taken into sodium citrate for platelet function, and for soluble P-selectin, soluble E-selectin, and von Willebrand factor (enzyme-linked immunosorbent assay, R&D System, Abingdon, UK and Dako, Ely, UK). hsCRP was measured by immunoturbimetry (Roche Diagnostics, Burgess Hill, UK). Monitoring the platelet responsiveness to clopidogrel was by the phosphorylation of intracellular regulatory protein—vasodilator-stimulated phosphoprotein (VASP), commonly used as marker of P2Y₁₂ receptor reactivity, and expression of CD62P and PAC-1 using flow cytometry^10,14,15,16 (Diagnostica Stago, Reading, UK). Platelet aggregation was performed on 270 μL aliquots of platelet-rich plasma (1,000 rpm, 10 minutes) with a four-channel PAP aggregometer (Alpha Laboratories, Basingstoke, UK) according to standard protocols using 30 μ agonists arachidonic acid (AA; Sigma Aldrich, Gillingham, Kent, UK, 0.5 mg/mL) and adenosine diphosphate (ADP; Sigma Aldrich, 10 and 20 μmol).¹⁶ Percentage of light transmission was noted 5 minutes after the addition of the agonist.

In whole blood flow cytometry (FACScalibor, Becton Dickinson, Oxford, UK), we assessed resting and activated platelets by increased expression of CD62P and PAC-1.^17,18 Briefly, the platelet cloud was identified in forward and side scatter, gated, and platelet identity confirmed by CD62P-APC and CD42a-PerCP (defining platelet gpIX) (both mAbs Becton Dickinson) in accordance with international guidelines.¹⁹ Specificity for % was set by isotype controls. Reaction tubes were incubated with ADP at 2 × 10⁻⁴ M for 2 minutes, AA at 125 μg/mL for 5 minutes, or saline before the addition of antibodies to detect the resting and activated expression of CD62P and PAC-1 (Becton Dickinson) as percentage of cells expressing each marker above isotype background, and for mean fluorescence intensity (MFI). The intra-assay coefficients of variation were < 5% for percentage marker expression and < 7% for MFI (n = 5 and 4 determinations, respectively).

Vascular function was assessed after 5 minutes of rest in the supine position. The SphygmoCor system (Artcor, Sidney, Australia) and a single high-fidelity applanation tonometer were used to measure pulse wave velocity (PWV).²⁰ Pulse waves were obtained sequentially from the carotid and radial artery, and from the carotid and femoral artery. The PWV was calculated from the transit time and the distance between these two arterial sites, determined in relation to the R-wave of the electrocardiogram, with patients lying in a horizontal position, and is reported in meters per second. Three complete sets of data were sampled and the average value was used as result. The SphygmoCor tonometer was placed over the right radial artery and three measures were collected, the result being the mean value. The device then analyzed the pulse wave using a validated generalized transfer function.

We hypothesized that a Spearman correlation coefficient (r) of 0.5 provides sufficient evidence of a significant pathophysiological effect. To achieve this at α = 0.05 and 1 − β = 0.8, n = 29 data points are demanded.²¹ We recruited 10% in excess of this figure for added confidence. This sample size provides the α = 0.05 power assurance that a two-sided change in a normally distributed index of 20% of a standard deviation is reliable. Data distributed normally are presented as mean and standard deviation, and were analyzed by paired t test. Data distributed nonnormally are presented as median and interquartile range and analyzed by Wilcoxon test. Categorical data were analyzed by chi-squared testing. A p < 0.05 was considered statistically significant.

Results

Table 1 shows flow cytometry data. The effect of cessation of clopidogrel was to increase the percentage of platelets expressing CD62P and the MFI after ADP stimulation. There was also an increase in the percentage of platelets expressing PAC-1 after ADP stimulation, and the PAC-1 MFI after both AA and ADP stimulation (summarized in Fig. 1). The median (interquartile range) VASP platelet reactivity index (reflecting clopidogrel action) was 36.7 (21.4 to 62.7) MFI units. This index failed to correlate significantly with any other index. The largest correlation coefficient obtained (r = 0.39, between the VASP platelet reactivity index and soluble P-selectin, although seemingly respectable, was markedly less than that required to minimize types 1 and 2 statistical error (r > 0.5 for a sample size of n = 32).

Table 1. Flow cytometry parameters of platelet function in patients on clopidogrel and after its withdrawal.

	On clopidogrel	Off clopidogrel	p value
% CD62P positive platelets at rest	17.6 (4.9)	15.8 (4.9)	0.143
% CD62P positive platelets after AA stimulation	22.8 (8.9)	25.0 (9.2)	0.202
% CD62P positive platelets after ADP stimulation	35.9 (11.7)	52.7 (16.5)	< 0.001
MFI of CD62P positive platelets at rest	29.8 (6.4)	29.6 (2.3)	0.807
MFI of CD62P positive platelets after AA stimulation	35.6 (9.4)	39.6 (11.6)	0.002
MFI of CD62P positive platelets after ADP stimulation	47.3 (13.5)	74.3 (30.6)	< 0.001
% PAC-1 positive platelets at rest	3.05 (2.0–5.6)	3.2 (1.9–6.3)	0.801
% PAC-1 positive platelets after AA stimulation	16.7 (9.1–19.7)	13.5 (6.3–21.6)	0.646
% PAC-1 positive platelets after ADP stimulation	35.2 (12.7)	49.3 (17.8)	0.006
MFI of PAC-1 positive platelets at rest	30.3 (2.3)	31.3 (3.2)	0.069
MFI of PAC-1 positive platelets after AA stimulation	36.9 (32.0–40.3)	39.5 (35.0–46.6)	0.001
MFI of PAC-1 positive platelets after ADP stimulation	56.3 (50.5–69.6)	94.6 (70.2–116.9)	< 0.001

Abbreviations: AA, arachidonic acid; ADP, adenosine diphosphate; MFI, mean fluorescence intensity.

Note: Date mean (standard deviation) or median (interquartile range).

Fig. 1.

Platelet responses to AA and ADP on clopidogrel and off clopidogrel. AA, arachidonic acid; ADP, adenosine diphosphate; MFI, mean fluorescence intensity.

There was no change in platelet aggregation to AA (as there was no change in aspirin use), but the response to both concentrations of ADP increased, reflecting cessation of clopidogrel therapy (Table 2). Soluble P-selectin increased upon cessation of clopidogrel, but there was no change in inflammatory (interleukin-6 and hsCRP) or vascular markers (von Willebrand factor, soluble E-selectin, or arterial stiffness). The change in soluble P-selectin failed to correlated significantly with the change in the number of platelets expressing membrane P-selectin induced by ADP (r = 0.018, p = 0.923), or the change in the MFI of P-selectin expression when stimulated by AA (r = − 0.163, p = 0.389) or by ADP (r = − 0.127, p = 0.496). There was no change in the mean arterial blood pressure after clopidogrel withdrawal (91 [13] vs. 93 [11] mm Hg), which also could partly explain the absence of changes in vascular wall characteristics.

Table 2. Platelet, inflammatory, and vascular responses to stopping clopidogrel.

	On clopidogrel	Off clopidogrel	p value
Platelet aggregation:
To arachidonic acid	17.5 (10.2–25.1)	18.0 (10.0–26.0)	0.666
To 10 μmol ADP	32.0 (15.5–41.7)	60.0 (51.0–66.0)	< 0.001
To 20 μmol ADP	39.0 (27.0–52.7)	69.0 (60.0–73.0)	< 0.001
Soluble P-selectin (ng/mL)	40.7 (8.4)	43.9 (10.7)	0.006
Interleukin-6 (pg/mL)	3.3 (1.1–20.8)	3.0 (1.6–23.0)	0.415
hsCRP (mg/mL)	1.0 (0.5–1.5)	1.1 (0.6–1.8)	0.454
von Willebrand factor (IU/dL)	95 (32)	89 (16)	0.284
Soluble E-selectin (ng/mL)	55 (45–86)	61 (48–84)	0.805
Pulse wave velocity (carotid–radial) (m/s)	7.4 (7.0–8.2)	7.1 (6.7–7.9)	0.149
Pulse wave velocity (carotid–femoral) (m/s)	8.7 (7.4–10.4)	8.3 (7.2–10.5)	0.126

Abbreviations: ADP, adenosine diphosphate; hsCRP, high sensitivity C-reactive protein.

Note: Data mean (standard deviation) or median (interquartile range), p value by paired t test or Wilcoxon test.

Discussion

Although the population value of clopidogrel in reducing thrombosis is clear, there may be nonplatelet effects, such as in inflammation,²² although this is controversial.^{7,8,9,10,11,12,13} Using a panel of plasma markers and physiological vascular function, we have been unable to detect any gross effect of clopidogrel use on the circulation or on inflammation. Some patients suffer thrombosis despite the supposed optimum antiplatelet therapy, implying lack of effect.^4,5 Indeed, our data underline the variability in the response of an individual's platelets, with the VASP platelet reactivity index varying 34-fold from a minimum of 2.3 MFI units to a maximum of 79.1 MFI units. Nevertheless, we found no major relationship between this index and any of the vascular and inflammatory markers.

In studies smaller than ours, recruiting 10 subjects, Holowatz et al⁸ found an effect of clopidogrel on reflex cutaneous vasodilatation, whereas França et al⁷ reported a significant correlation (r = 0.48, p = 0.025) between the plasma concentration of clopidogrel and the number of endothelial progenitor cells in 26 patients. Waehre et al¹² noted that clopidogrel significantly up-regulated the gene expression of inflammatory markers in peripheral blood mononuclear cells in 18 patients. As our sample size was greater, we conclude that the methods used by our colleagues^7,8,12 may be more sensitive in assessing the effects of this antiplatelet. Immediate cessation of clopidogrel is not associated with any thrombotic events.²³

Like Wykrzykowska et al,¹³ we used a cessation model, and like them also found an increase in soluble P-selectin. However, their larger study (n = 98 patients) also found a reduction in hsCRP, which we were unable to confirm, possibly because our sample size was one-third of their sample size. P-selectin is a component of the membrane of the α granule, so that its appearance at the surface implies degranulation and so platelet activation, and this explains increased expression after agonist stimulation. Soluble P-selectin reflects that membrane component that is actively shed or is cleaved by plasma proteinases.^24,25 The finding that soluble P-selectin rises after clopidogrel cessation implies an ADP-related pathway to α degranulation that is present despite the suppression of the cyclo-oxygenase pathway by aspirin. However, it could be argued that the cyclo-oxygenase pathway is not completely inhibited as platelets still responded to AA by expressing increased P-selectin (Fig. 1). Indeed, platelet responses to AA were higher when off clopidogrel (58% increase in number of cells expressing P-selectin, 34% increase in MFI) than when on clopidogrel (29 and 20%, respectively), indicating that there is cross-talk between ADP and AA pathways. Increased responsiveness of platelets to AA despite aspirin therapy may reflect a degree of aspirin resistance.⁶ However, the magnitude of the change in soluble P-selectin failed to correlate with the change in membrane expression, perhaps suggesting that the cleavage of P-selectin from the membrane to produce the soluble form is nonspecific.

The increase in arterial stiffness is a long-term process which develops as a result of different mechanisms: (1) a breakdown of the elastic structure (elastin fibers) in the arterial walls which is the prime cause of increasing stiffness in the aorta with aging; (2) damage to the endothelium/smooth muscle mechanism by which arterial stiffness is dynamically controlled (this is the prime cause of arterial stiffening in the muscular conduit arteries); (3) an elevation in mean arterial pressure which increases the stiffness of an artery.²⁶ It is therefore possible that the vascular wall needs more than 2 weeks to recover from the effects of clopidogrel (if any).

We hypothesized a positive effect of clopidogrel on endothelial function, expecting an improvement in carotid–radial and carotid–femoral PWV (predominantly muscular-type arteries), and in levels of von Willebrand factor and soluble E-selectin. However, as our data failed to support this hypothesis, we conclude that 75 mg clopidogrel daily has no major effect on vascular function or on plasma markers of inflammation. Furthermore, our data also point to continued responsiveness of platelets to AA and ADP despite the use of agents designed to prevent this. A corollary is that doses of aspirin and clopidogrel may need to be increased if additional platelet suppression is desired.