Lipophilic statins interfere with the inhibitory effects of clopidogrel on platelet function — a flow cytometry study

Abstract

Aims Clopidogrel is a pro-drug which is converted to an active, unstable drug by cytochrome P450 (CYP). The active drug irreversibly blocks one specific platelet adenosine 5′-diphosphate (ADP) receptor (P2Y12). It has been recently suggested that the most abundant human CYP isoform, 3A4, activates clopidogrel. Since certain lipophilic statins (i.e. simvastatin, atorvastatin, lovastatin) are a substrate of CYP3A4, we were interested in potential drug interactions between clopidogrel and statins.

Methods In patients with coronary artery disease (n=47) in whom clopidogrel treatment was initiated for balloon angioplasty and stent implantation, blood samples were taken at 0, 5 and 48h after oral administration of clopidogrel (loading dose 300mg, followed by 75mg daily). ADP-stimulated (1, 10, 100μmol/l) expression of P-selectin (CD62P) on platelets was measured by flow cytometry, and used as a marker for the antiplatelet effect of clopidogrel.

Results Pre-treatment with statins (atorvastatin, simvastatin) reduced significantly (10μmol/l ADP stimulation) the inhibitory effects of clopidogrel during the loading phase (relative reduction after 5h 29.3%) and, to a lesser extent during the maintenance phase (relative reduction after 48h 16.6%). In addition we found a considerable individual heterogeneity in the response and three patients (6%) were identified in whom clopidogrel exerted almost no effect.

Conclusion Certain statins which are substrates of the CYP3A4 isoform competitively inhibit the metabolic activation of clopidogrel. As a result the relative clopidogrel induced platelet inhibition (P-selectin-expression) is diminished—but still there is a relative clopidogrel effect of more than 80% in the maintenance phase. It may be reasonable to test the therapeutic efficacy of clopidogrel in those patients who require long-term treatment.

1 Introduction

Clopidogrel is an orally administered thienopyridine agent that selectively and irreversibly inhibits ADP-induced platelet aggregation. The drug is inactive in vitro and requires in vivo oxidation by hepatic/intestinal cytochrome P450 isoenzymes. The highly unstable active metabolite of clopidogrel presumably forms a disulphide bridge between a reactive thiol group and a cysteine residue of the P2Y12 platelet receptor.¹Pharmacodynamic investigations emphasized the need of aclopidogrel loading dose of 300mg during the initial phase followed by once daily doses of 75mg.²

As mentioned above, the antiaggregating activity of clopidogrel requires hepatic/intestinal activation by cytochrome P (CYP) 450.³In animals, the efficacy of clopidogrel is increased after pre-treatment with 3-methylcholanthrene and β-naphthoflavone, indicating that a CYP1A subfamily pathway activates clopidogrel.⁴The use of specific antibodies directed against the various CYP subfamilies confirmed this suggestion.⁵In contrast, in genetically engineered human microsomes containing a single human P450 isoenzyme, CYP3A4 and 3A5 appear primarily responsible for clopidogrel metabolism.⁶The identification of CYP3A4 as the crucial enzyme for activation of the prodrug clopidogrel could be of clinical importance since patients with coronary artery disease are frequently treated with statins. Statins with the exception of pravastatin are also metabolized by certain CYP450 subfamilies including CYP3A4, and could potentially interfere with the hepatic/intestinal activation of clopidogrel. Very recently, Lau and co-workers⁷reported on 19 patients undergoing coronary artery stenting that atorvastatin, a substrate of the CYP3A4 subfamily, impairs the antiplatelet effect of clopidogrel in a dose-dependent manner. In this study, the effect of clopidogrel was measured by an electronic impedance cell counting technique. Because this study has potential limitations (e.g. patients were pre-treated with the glycoprotein IIb/IIIa inhibitor eptifibatide before clopidogrel application, not well established point of care platelet test using not yet standardized conditions), we thought to re-evaluate this important question using an alternative and established technique to measure platelet activation (platelet CD62P expression). The necessity for further studies is emphasized by previous contrary findings. The re-analysation of data from the PRONTO Trial (Plavix Reduction of New Thrombus Occurrence)⁸could not disclose an inhibitory effect of various statins on the clopidogrel activity in 25 patients.

In the present prospective study, we measured the ADP-induced P-selectin (CD62P) expression on platelets before and after administration of clopidogrel in the presence and absence of statins in patients undergoing coronary angioplasty. In contrast to previous studies, we used various ADP concentrations, and included only patients who had been on lipid lowering therapy for at least 1 week. Moreover, we studied the platelet inhibition of clopidogrel early (5h) and late (48h) after administration of the loading dose.

2 Methods

2.1 Patients

This study included 47 patients who received clopidogrel (loading dose 300mg p.o., 75mg p.o. the following days) for elective PTCA and stent placement. Group A (control group: n=22, 13 male, nine female, mean age 65±15 years, body mass index 26.5±2.9kg/m²) either did not receive lipid lowering drugs, or had been off lipid lowering drugs for more than 3 months. Group B (statin group: n=25; 16 male, nine female, mean age 63±14 years, body mass index 27.0±2.5kg/m²) was on statin treatment for at least 1 week, and 22 patients out of this group were treated with statins for more than 3 months. The statins used were (number of patients treated): atorvastatin 20–40mg/day (n=17), and simvastatin 10–20mg/day (n=8). The patients enrolled had no history of bleeding diathesis, no history of drug or alcohol abuse or liver disease, a normal prothrombin time, a normal platelet count, and a serum creatinine <1.2mg/dl. The clopidogrel loading dose was given a day prior to PTCA/stenting. All patients were on long-term aspirin (100mg/dl), and received intravenous heparin during stenting to achieve an activated clotting time >300s. Glycoprotein IIb/IIIa inhibitors were not given at all. The CYP pathway relevant co-medication (potential substrate, inducer or inhibitor of CYP isoforms) consisted of (number of patients in group A/B): digitoxin (1/1); verapamil (2/1), diltiazem (1/2) and steroids (1/0). The patients provided informed consent and this study was approved by the local ethics committee.

2.2 Blood sampling and flow cytometry

Peripheral venous blood (5ml, after the first 5ml was dispatched) was drawn with a 21G needle into blood collecting tubes containing sodium citrate (0.5ml), and gently mixed 3–5 times. Blood samples were drawn before clopidogrel loading, and at 5h and 48h post loading. Usually, the blood samples at 0h and 5h were drawn the day before, and at 48h the day after successful PTCA procedure. Platelet immunostaining was performed as previously described.^9,10For basal measurements, a 0.5ml specimen of the blood-citrate mixture was mixed with 0.5ml paraformaldehyde (1%) and fixed for 15min at room temperature. The fixed and unfixed specimens (1ml) were centrifuged (5000rpm, 45s) to obtain platelet-rich plasma (PRP).¹¹PRP from unfixed samples was stimulated with adenosine 5′-diphosphate (ADP; final concentrations 1, 10 and 100μmol/l) for 15min at room temperature. Thereafter, samples (20μl) were incubated with monoclonal antibodies (5μl) (fluorescein isothiocyanate (FITC)-conjugated anti-CD62P and FITC-conjugated IgG₁) for 15min at room temperature (samples were kept away from light). Incubation was stopped by adding 500μl phosphate-buffered saline. The surface expression of CD62 receptors (P-selectin) was determined by flow cytometry (Beckman Coulter Epics XL, Krefeld, Germany). Platelets were identified and gated by forward light and side-scatter profiles, 30 000 platelets were analysed at a flow rate <1000platelets/s. The degree of platelet activation was measured as mean FITC-fluorescence of anti-CD62P bound to the platelets (in arbitrary fluorescence units; MnX) minus the FITC-fluorescence of anti-IgG bound. The time interval between blood sampling and end of flow cytometry measurements was <60min.

2.3 Materials

The antibodies FITC-conjugated anti-CD62P and mouse FITC-isotype IgG₁were purchased from Beckman Coulter (Krefeld, Germany); phosphate-buffered saline was from Dulbecco (Biochrom, Berlin); paraformaldehyde and adenosine 5′ diphosphate were from Sigma Chemicals.

2.4 Statistics

All values are reported as mean±SD. The number of patients is represented by n. All measurements were taken in duplicate. Inhibition of ADP-induced CD62P expression by clopidogrel treatment was expressed as the percent difference of mean fluorescence before and after clopidogrel application. The differences between baseline and post-treatment values were analysed with the paired two-sample t-test. Comparisons between groups were made with the unpaired two-sample test. Statistical significance was set at P<0.05.

3 Results

Basal expression of CD62 on platelet surface in fixed blood specimens was low (and unchanged compared to the unfixed and not stimulated specimen), indicating negligible activation of platelets during blood sampling. Fig. 1summarizes the mean fluorescence (MnX) of the ADP-stimulated CD62 expression on platelets for group A (control group) and group B (statin group) before, 5h and 48h after, clopidogrel administration. Platelet activation by ADP induced a strong concentration-dependent increase in the surface expression of P-selectin, which was of similar magnitude in both groups (P=ns).

The ADP-inducible P-selectin expression was clearly reproduced in both groups at 5h and 48h following the clopidogrel loading dose. The magnitude of this inhibitory effect of clopidogrel was somewhat influenced by the concentration of ADP used for P-selectin stimulation, thus, all results are shown for the various ADP concentrations (Table 1).

The inhibitory effects of clopidogrel were impaired in the statin group as compared to the control group, these differences were best seen when 10μmol/l ADP wasused for platelet stimulation. The attenuation of the clopidogrel effects were more pronounced during the early loading phase (relative reduction of approximately 29% as compared to the corresponding values in the control group) as during the maintenance phase (relative reduction of approximately 16%). Fig. 2shows a side-by-side comparison of the clopidogrel effects on platelet function in the control and statin group. When 100μmol/l ADP was used for platelet stimulation, the interference of statins with the inhibitory effects of clopidogrel on P-selectin expression was much less impressive: significant differences were only observed for the whole statin group during the early loading phase (Table 1).

A subanalysis of the different statins and dosages used (20/40mg atorvastatin, 10/20mg simvastatin)revealed only trends, but no significant differences in their inhibitory effects on clopidogrel metabolism(Table 1).

The effect of clopidogrel on ADP-induced P-selectin expression varied widely. In fact, three patients (6%) were identified (one of group A, 2 of group B receiving 10mg simvastatin and 20mg atorvastatin, respectively), in whom clopidogrel had exerted no inhibitory effect at all after 5h and 48h.

4 Discussion

In mammalian species, CYP gene subfamilies are expressed as microsomal monoxygenases and include many different isozymes. More than 150 isoforms have been identified up until now. These represent principally the CYP1 family, including CYP1A1 and CYP1A2 genes which can be induced by polycyclic aromatic hydrocarbons, β-naphthoflavone, and 3-methylcholanthrene, the CYP2 family, including phenobarbital—(CYP2B and CYP2C subfamilies) and ethanol- or isoniazid-inducible genes (CYP2E subfamily including CYP2E1), and the CYP3 family, whose gene is inducible by macrolide antibiotics such as rifampicin and glucocorticoids (CYP3A1 to CYP3A4).

Statins are 3-hydroxy-3 methylglutaryl coenzyme A reductase inhibitors. Most of the lipophilic statins e.g. lovastatin, simvastatin and atorvastatin are substrates of CYP subfamilies, and are metabolized mainly by CYP3A4, fluvastatin mainly by CYP2C9. Pravastatin is the only hydrophilic HMG-CoA reductase inhibitor that is eliminated unchanged.¹²Whenever a medication (substrate) is metabolized by the CYP system and is taken simultaneously with an agent that decreases the activity of the same enzyme system (i.e. inhibitor), the result will be an increase in the concentration of this substrate, which in turn will increase the potential for adverse drug reactions. This scenario has been extensively studied and reviewed for statin-associated myopathy and rhabdomyolysis.^13–15On the other hand, little is known on whether medication with statins may interfere vice versa with the CYP-dependent metabolic activation of a pro-drug, such as clopidogrel. This issue is of potential interest since both, statins and clopidogrel are widely used as a co-medication in patients with coronary artery disease undergoing intracoronary repair and stent implantation and in patients with acute coronary syndromes.

The present study demonstrates that atorvastatin and simvastatin interfere with the inhibitory effects of clopidogrel on platelet function under certain conditions. Most of the significant results were seen in our study when a moderate concentration of ADP (10μmol/l) was used for platelet stimulation. The impact of statins on the inhibitory effect of clopidogrel on platelet function, however, was much less impressive when maximal concentrations of ADP were used (100μmol/l). We can only speculate that very high concentrations of ADP stimulate platelets not only by specific P2Y12 receptors but also by more unspecific sites. Subsequently, any changes in the plasma concentration of the active clopidogrel metabolite (and thus blockade of P2Y12 receptors) are probably less detectable under these assay conditions. Furthermore, we did not notice a significant interference between clopidogrel and statins when 1μmol/l ADP was used for platelet stimulation; ADP at this low concentration exerted only a minor increase in P-selectin expression (Fig. 1), thereby the assay conditions may be suboptimal to disclose any changes on P2Y12 receptor occupation.

In contrast to the study by Lau and coworkers,⁷the inhibitory effects of statins on the ability of clopidogrel to induce platelet aggregation were much less pronounced in our study. Lau and coworkers reported on a subgroup of five patients that atorvastatin in a daily dose of 40mg ‘completely’ inhibited the antiplatelet activity of clopidogrel. In our hands, the inhibitory effects of clopidogrel on the ADP-mediated P-selectin expression dropped during the loading (5h) and maintenance phase (48h) by about 29% and 16%, respectively, in thepresence of statins. The interference between statins and clopidogrel activation was more evident during the loading phase of clopidogrel, but to a lesser extent during the maintenance phase. As a trend, we noted at higher doses of statins a stronger interference on the clopidogrel efficacy, but these differences did not reach significant levels (Table 1). It appears that the variability in the individual response to clopidogrel is higher than the (minor) impact of different doses of statins on clopidogrel efficacy. This difference to the data from Lau and coworkers is not clear, and may be related to the different methods used for the quantification of platelet function (aggregation vs P-selectin expression, different concentrations of ADP used for platelet stimulation, no glycoprotein IIb/IIIa inhibitor pretreatment in our study). Furthermore, it could be related to a high variability of the inhibitory effect of clopidogrel, as demonstrated in this study as well as being reported previously by others.¹⁶

An interesting observation in our study group is that three of 47 patients (6%) apparently did not respond to clopidogrel therapy at all, at least within 48h of drug application. The reason for this phenomenon is not known, and the disclosure of the underlying pathomechanism was beyond the scope of this study. This phenomenon is unlikely to be related to a congenital defect of the platelet ADP receptor,¹⁷since P-selectin expression was still inducible with ADP. This phenomenon could be related to highly variable interindividual CYP-expression and/or activity.^18,19It has been reported that the expression of CYP3A4 varies 40-fold in individual human livers, and metabolism of CYP3A4 substrates varies at least 10-fold in vivo.²⁰A large-scale study will be necessary to define the frequency of potential clopidogrel non-responders.

4.1 Clinical implications

The clinical impact of our observation is not known. The present study design does not allow an association between the magnitude of the antiaggregating effect of clopidogrel and the rate of cardiac events during follow-up. A very recent study,²¹however, demonstrates a higher rate of cardiovascular death in patients on clopidogrel treatment as compared with those on ticlopidine (which does not require conversion by CYP3A4).²²In this study, 86% of patients were on statins.

The inhibitory effect of certain statins on the clopidogrel bio-availability may be neutralized either by an earlier administration of the loading dose before any planned interventions or by increasing the loading dose. A recent study demonstrated that the application of 450mg instead of 300mg clopidogrel as the loading dose shortened the period until the maximum effect was achieved.²³Whether an increase in the loading dose may be also effective in possible non-responders, is not known. At least, it appears reasonable to test the therapeutic efficacy of clopidogrel in patients before they will require long-term treatment (e.g. patients undergoing intracoronary brachytherapy after stent implantation).

4.2 Limitations

This non-randomized study is limited by its observational nature, and the relative low number of patients included. Unadjusted tests were done to compare the groups, and thus the observed differences in the inhibitory effects of clopidogrel cannot be necessarily attributed to the statin pretreatment.

We thank A. A. Weber, PhD from the Institute of Pharmacology and Clinical Pharmacology, Heinrich-Heine-University, Düsseldorf, Germany, for his kind advice and support concerning the method.

References