A Meta-Analysis of Impact of Proton Pump Inhibitors on Antiplatelet Effect of Clopidogrel


  • M. Chen and J.F. Wei contributed equally to this investigation.

D.J. Huang, Department of Cardiology, West China Hospital, Sichuan University, 37 Guoxue Street, Chengdu, 610041, PR China.
Tel.: 862885423362;
Fax: 862885422169;
E-mail: huangdjmd@yahoo.com


Previous mechanistic studies have suggested a possible interaction between proton pump inhibitor (PPIs) and clopidogrel. However, the results of clinical trials about the effects of PPIs on safety and efficacy of clopidogrel are controversial. The study sought to estimate the impact of PPIs on antiplatelet effect of clopidogrel. The study performed a meta-analysis of comparative concomitant use of clopidogrel with PPIs versus clopidogrel without PPIs studies published or presented to October 2010. Cardiovascular death, readmission for myocardial infarction/readmission for acute coronary syndrome, and nonfatal stroke were set as clinical endpoints. In randomized control trials (RCTs), the clinical endpoints risk ratio for clopidogrel with PPIs versus clopidogrel without PPIs was 1.20 (P= 0.34) in the random-effects model and 1.03 (P= 0.63) in the fixed-effects model. In observational studies, the risk ratio for the clinical endpoints for clopidogrel with PPI versus clopidogrel without PPI was 1.40 (P < 0.001) in the random-effects model and 1.49 (P < 0.001) in the fixed-effects model. Different assay methods showed that coadministration of clopidogrel with PPIs was associated with attenuation of clopidogrel's antiplatelet effect in vitro. This meta-analysis indicated an obvious discrepancy between RCTs and observational studies with respect to the interaction between PPIs and clopidogrel.


Clopidogrel in combination with aspirin (dual antiplatelet therapy) is a cornerstone of therapy for prevention of ischaemic events and improving outcomes following acute coronary syndrome (ACS) and percutaneous coronary intervention (PCI) with stenting [1]. The current American College of Cardiology/American College of Gastroenterology/American Heart Association guideline recommends that patients receiving dual antiplatelet therapy with clopidogrel and aspirin receive a proton pump inhibitor (PPI) to reduce the risk of gastric side-effects [2]. Approximately half of patients taking dual antiplatelet therapy were prescribed with a PPI in real world practice [3]. Clopidogrel is a pro-drug that is converted by the hepatic cytochrome P450 (CYP450) system, and the active compound irreversibly inhibits the ADP P2Y12 receptor on circulating platelets. The need for in vivo bioactivation makes clopidogrel vulnerable to drug–drug interactions with other drugs, such as PPIs, metabolized by hepatic CYP450 isoenzymes.

Many mechanistic studies suggest that combined use of clopidogrel and PPIs may attenuate antiplatelet effects of clopidogrel [4–7], and the data from large scale observational studies have reported that concomitant use of clopidogrel and PPIs after hospital discharge was associated with an increased risk of adverse outcomes than use of clopidogrel without PPIs [8,9]. Based on these data, both the EMEA (European Medicines Agency) and the FDA (US Food and Drug Administration) have issued public statements on a possible adverse interaction between clopidogrel and PPIs [10,11]. However, results from the COGENT trial comparing omeprazole versus placebo in patients taking clopidogrel, has found no difference in the risk of cardiovascular events in patients taking omeprazole [12]. To address this knowledge gap and provide a summary of the current available evidence, we undertook the present meta-analysis.


Criteria for Inclusion/Exclusion

Studies eligible for inclusion included those published or presented at a major cardiovascular meeting to October 2010 comparing combination of clopidogrel and PPIs with clopidogrel alone.

Data Search and Acquisition

Several sources were searched for published/presented studies, including MEDLINE, the Cochrane database, EMBASE, http://www.clinicaltrials.gov, http://www.clinicaltrialresults.org, http://www.tctmd.com, http://www.cardiosource.com, and abstracts/presentations from major cardiovascular meetings.

Study Endpoints

The clinical endpoints chosen for this analysis were cardiovascular death, readmission for MI (myocardial infarction)/readmission for ACS (MI or unstable angina) and nonfatal stroke.

Data Extraction

Microsoft Excel and Review Manager version 5.0 for Windows were used for data extraction. Data extraction and study selection were performed by two independent investigators (J.F.W and Y.N.X). Disagreements regarding the appropriateness of selected studies included for analysis were resolved by discussion and consultation with other investigators of our group.

Statistical Analysis

Review Manager version 5.0 for Windows was utilized to analyze clinical endpoints. Separate analyses of RCTs, including post hoc analysis of RCTs, and observational studies were prespecified, considering the inherent differences between these study designs. Dichotomous data was expressed as risk ratios (RR) with 95% confidence intervals. Both fixed-effects and random-effects models are reported. Statistical heterogeneity was assessed via chi-squared test and was quantified with the I-squared test.


Description of Studies

A total of 13 studies met the study criteria (Table 1). The data of two published abstracts derived from a retrospective cohort study (Clopidogrel Medco Outcomes study) with different number of subjects (16,690 vs. 14,383) [9,13]. One post hoc study [7] analyzed the data from two different previous RCTs (the Prasugrel in Comparison to Clopidogrel for Inhibition of Platelet Activation and Aggregation-Thrombolysis in Myocardial Infarction 44 trial (PRINCIPLE-TIMI 44 trial)[14] and Trial to Assess Improvement in Therapeutic Outcomes by Optimizing Platelet Inhibition with Prasugrel–Thrombolysis in Myocardial Infarction (TRITON-TIMI) 38 trial)[15], and in this review the findings from this post hoc study were considered as two different RCTs. Most participants were undergoing PCI due to stable angina or ACS.

Table 1. Characteristics of included studies
AuthorsDesignPatientsTotalInterventionsFollow upEndpointsDefinition of hyporesponsivenessResultsSponsor
  1. RCT, Randomized controlled trial; RCS, Retrospective cohort study; COS, Cross-sectional observational study; NCS, Nested case-control study; PCI, Percutaneous coronary interventions; AMI, Acute myocardial infarction; MI, myocardial infarction; ACS, Acute coronary syndrome; PPI–, without proton pump inhibitor; PPI+, with proton pump inhibitor; VASP, Vasodilator-stimulated phosphoprotein phosphorylation; MEA, Multiple electrode platelet aggregometry; PRI, Platelet reactivity index; IPA, Inhibition of platelet aggregation; LTA, light-transmission aggregometry; CHR, Clopidogrel hyporesponsiveness; PRU, P2Y12 reaction unit; RPR, Residual platelet reactivity; N/A, Not applicable; HR, Hazard ratio; OR, Odd Ratio; vs, versus.

Trials to investigate clinical endpoints
Randomized control studies
Bhatt et al. 2010 [12]RCTAMI, unstable angina or after stent implantation3761Clopidogrel with/
without Omeprazole
106 daysCardiovascular death, nonfatal MI, revascularization, or strokeN/AOmeprazole+vs Omeprazole- (HR, 0.99; 95% CI 0.68–1.44)Industry
O'Donoghue et al. 2009 [7] Posthoc analysisPost-PCI13608Clopidogrel with/without PPIs15 monthsCardiovascular death, MI, or stroke;N/APPI+vs PPIP=0.72Industry
Dunn et al. 2008 [16] Post hoc analysisPost-PCI1053Clopidogrel with/without PPIs1 yearDeath, MI, strokeN/APPI+vs PPI– (OR, 1.633; 95%CI 1.015–2.627)Investigator
Observational studies
Ho et al. 2009 [8]RCSAMI or unstable angina8205Clopidogrel with/without PPIs521 daysDeath or readmission for ACSN/APPI+vs PPI–P < 0.00001Investigator
Stanek et al. 2009 [9]RCSPost-PCI16,690Clopidogrel with/without PPIs1 yearCardiovascular death, nonfatal MI, revascularization, or strokeN/APPI+vs PPI– (HR, 1.51; 95%CI 1.39–1.64)Investigator
Charlot et al. 2010 [17]RCSAMI24,702Clopidogrel with/without PPIs1 yearCardiovascular death, nonfatal MI, or strokeN/APPI+vs PPI- (HR, 1.29; 95%CI 1.17–1.42)Investigator
Ray et al. 2010 [18]RCSPost-PCI10,152Clopidogrel with/
without PPIs
1 yearCardiovascular death, nonfatal MI, or strokeN/APPI+vs PPI- (HR, 1.01; 95%CI 0.76–1.34)Investigator
Aubert et al. 2008         
[13]RCSPost-PCI14,383Clopidogrel with/
without PPIs
1 yearCardiovascular death, nonfatal MI, revascularization, or strokeN/APPI+vs PPI–; P < 0.0001Investigator
Juurlink et al. 2009 [19]NCSAMI2791Clopidogrel with/without current PPIs90 daysReadmission for ACSN/APPI+vs PPI- (OR, 1.27; 95%CI 1.03–1.57)Investigator
Trials to investigate platelet function
Gilard et al. 2008 [4]RCTPost-PCI140Clopidogrel with/without Omeprazole7 daysPRI assessed by VASPCHR was defined as PRI > 50%. PRI Omeprazole+vs Omeprazole- P < 0.001Investigator
        CHR Omeprazole+vs Omeprazole- P = 0.0005 
Price et al. 2009 [5]COSPatients receiving Clo >7 days or 300 mg loading dose377Clopidogrel with/without current PPIsN/ARPR assessed by VerifyNow P2Y12 assayCHR was defined as high RPR (≥235 PRU) RPR PPI+vs PPI–; P = 0.02Investigator
        CHR PPI+vs PPI–; P = 0.06 
Sibbing et al. 2009 [6]COSPost-PCI1000Clopidogrel with/without PPIsN/APlatelet aggregation assessed by MEACHR was defined as platelet aggregation higher than the upper quintile (20%) of study population. Platelet aggregation Omeprazole+vs placebo P = 0.001; Esomeprazole+vs placebo P = 0.88; Pantoprazole+vs placebo P = 0.69;Investigator
        CHR Omeprazole+vs placebo P = 0.008 Esomeprazole+vs placebo P = 0.87; Pantoprazole+vs placebo P = 0.76 
O'Donoghue et al. 2009 [7] Post hoc analysisPost-PCI201Clopidogrel with/without PPIs15 daysIPA assessed by LTACHR was defined as less than 20% IPA. IPA PPI+vs PPI–P < 0.05;Investigator
        CHR PPI+vs PPI–P = 0.005 

Cardiovascular Death, Readmission for MI/ACS and Nonfatal Stroke

Clinical endpoints were analyzed for a total of 3 RCTs (one RCT and two post hoc studies of RCTs) and four observational studies. There was some significant heterogeneity across RCTs (I2 67%, P= 0.05) and significant heterogeneity between observational studies (I2 96%, P < 0.001), suggesting that one should interpret summary estimates with caution. For these three RCTs, the RR for the clinical endpoints for clopidogrel with PPI versus clopidogrel without PPI was 1.20 (95% CI: 0.82–1.76, P= 0.34) in the random-effects model (Figure 1A) and 1.03 (95% CI: 0.91–1.18, P= 0.63) in the fixed-effects model (Figure 1B) [7,12,16]. For observational studies, the RR for the clinical endpoints for clopidogrel with PPI versus clopidogrel without PPI was 1.40 (95% CI, 1.15–1.70; P < 0.001) in the random-effects model (Figure 1C) and was 1.49 (95% CI, 1.43–1.55; P < 0.001) in the fixed-effects model (Figure 1D) [8,9,17,18]. A nested case-control study demonstrated that current use of PPIs in patients prescribed clopidogrel was associated with increased risk of reinfarction–adjusted odds ratio of 1.27 (95% CI, 1.03–1.57) [19].

Figure 1.

The effects of combination use of clopidogrel and PPI on clinical endpoints (cardiovascular death, readmission for MI/readmission for ACS and nonfatal stroke) in randomized control trials (random-effects model (A) and fixed-effects model (B)) and observational trials (random-effects model (C) and fixed-effects model (D)). PPI, proton pump inhibitor.


In RCTs, combination use of clopidogrel and PPIs compared with use of clopidogrel alone was associated with no detectable difference in cardiovascular events. From a review standpoint of observational studies, the nonrandomized coadministration of clopidogrel with PPIs versus without PPIs was associated with approximately 50% increased risk of cardiovascular death, readmission for MI/ACS, and nonfatal stroke. With respect to platelet function, coprescribing PPIs likely attenuated the antiplatelet effect of clopidogrel assessed by various different methods in vitro.

The Impact of PPIs on Anti-platelet Effects of Clopidogrel

In general, caution should be used when interpreting the summary estimates from the analysis owing to the significant heterogeneity between selected studies. The analysis of RCTs demonstrates that there are no interactions between administration of PPIs and clopidogrel in cardiovascular events, but in observational studies, concomitant use of clopidogrel and PPIs is associated with higher prevalence of cardiovascular events. There may be several explanations for the conflicting results between RCTs and observational studies. First, the limitations inherent in RCTs and observational studies may affect the clinical context and interpretation of results. Although the design of RCT can minimize the influence of various confounders and treatment selection bias, the size of RCTs employed in this clinical scenario are small and not adequately powered to evaluate hard clinical endpoints. In contrast to RCTs, selection bias can confuse interpretation of nonrandomized trials’ results. However, the large-scale observational studies may be more generalizable and less subject to enrollment bias, and always have enough power to assess the hard clinical endpoints. Thus, both types of data are useful in assessing the relative safety of coadministration of PPIs and clopidogrel. Second, patients prescribed clopidogrel in addition to PPIs are generally older and in poorer health compared to those prescribed clopidogrel alone in clinical practice [8,9], and the prognoses of these patients are comparatively worse. The design of observational study cannot eliminate these confounding factors when interpreting the trial results. Third, the COGENT trial [12] was stopped prematurely after the trial sponsor declared bankruptcy. The mean follow-up of this study was only 106 days, whereas, in either of the observational studies, the median follow-up was greater than one year [8,9,17,18]. The shorter follow-up intervals and comparatively smaller scale may underestimate the difference of recurrent rate of cardiovascular events between concomitant use of clopidogrel with PPIs and clopidogrel alone. Accordingly, reliance on the results of either RCTs or observational studies available so far should not be absolute, and due to be the evidence with highest grade, larger scale RCTs with longer follow up are needed to assess the interactions of PPIs and clopidogrel.

Although nonstandard measurements have been universally accepted, analyses of platelet function with different assay methods consistently demonstrate that PPIs can attenuate the antiplatelet effect of clopidogrel in vitro (Table 1) [4–7]. However, it is controversial whether the attenuation of the antiplatelet effect of clopidogrel might have been sufficient to translate into an increased risk of adverse outcomes. We should remain vigilant to this possibility due to previously reported interactions between atorvastatin and clopidogrel. It has been reported that atorvastatin may attenuate platelet inhibition by clopidogrel, however, such in vitro results are not associated with worse clinical outcomes [20–23]. Therefore due to the limitation of ex-vivo study, it is prudent to avoid using surrogate endpoints as a substitute of clinical events.


Meta-analysis is a powerful tool to summarize the results of multiple studies. However, due to inherent limitations (such as publication bias, heterogeneity of enrolled studies, and different populations), meta-analysis is no substitute for an adequately designed RCT and should be considered hypothesis generating [24]. Moreover, there are few high quality trials to investigate the possible interaction of PPIs and clopidogrel, thus preventing a comprehensive meta-analysis with respect to the issue at hand. Therefore, additional studies (especially large-scale RCTs) are clearly warranted in order to assess the safety and effectiveness of clopidogrel when used concurrently with PPIs.


This meta-analysis indicated an obvious discrepancy between RCTs and observational studies with respect to the interaction between PPIs and clopidogrel. Large-scale RCTs are required to ascertain the true effects of PPIs on clopidogrel with respect to cardiovascular events.


The work was supported by grants from the Chinese National Nature Science Foundation (grant numbers 30600607 and 30770877, Beijing, China), the National High-tech Research and Development Program of China (2006AA02A406, Beijing, China), and Chinese Postdoctoral Science Foundation (20060400305, Beijing, China).

The funding organizations had no role in the study design, data collection, data analysis, manuscript preparation, and/or publication decisions.

Conflict of interests

The authors have no conflict of interest.