ACCF/ACG/AHA 2010 Expert Consensus Document on
the Concomitant Use of Proton Pump Inhibitors and
Thienopyridines: A Focused Update of the ACCF/ACG/AHA
2008 Expert Consensus Document on Reducing the
Gastrointestinal Risks of Antiplatelet Therapy and NSAID Use
A Report of the American College of Cardiology Foundation Task Force on
Expert Consensus Documents
Writing
Committee
Members
Neena S. Abraham, MD, FACG, Chair*
Mark A. Hlatky, MD, FACC, FAHA,
Vice Chair†
Elliott M. Antman, MD, FACC, FAHA‡
Deepak L. Bhatt, MD, MPH, FACC, FAHA†
David J. Bjorkman MD, MSPH, FACG*
Craig B. Clark, DO, FACC, FAHA†
Curt D. Furberg, MD, PHD, FAHA‡
David A. Johnson, MD, FACG*
Charles J. Kahi, MD, MSC, FACG*
Loren Laine, MD, FACG*
Kenneth W. Mahaffey, MD, FACC†
Eamonn M. Quigley, MD, FACG*
James Scheiman, MD, FACG*
Laurence S. Sperling, MD, FACC, FAHA‡
Gordon F. Tomaselli, MD, FACC, FAHA‡
*American College of Gastroenterology Representative; †American
College of Cardiology Foundation Representative; ‡American Heart
Association Representative
ACCF
Task Force
Members
Robert A. Harrington, MD, FACC, Chair
Eric R. Bates, MD, FACC
Deepak L. Bhatt, MD, MPH, FACC
Victor A. Ferrari, MD, FACC
John D. Fisher, MD, FACC
Timothy J. Gardner, MD, FACC
Federico Gentile, MD, FACC
Mark A. Hlatky, MD, FACC
Alice K. Jacobs, MD, FACC
Sanjay Kaul, MBBS, FACC
David J. Moliterno, MD, FACC
Howard H. Weitz, MD, FACC
Deborah J. Wesley, RN, BSN
This document was approved by the American College of Cardiology Foundation
(ACCF) Board of Trustees in September 2010, by the American College of
Gastroenterology (ACG) Board of Trustees in September 2010, and by the American
Heart Association (AHA) Science Advisory and Coordinating Committee in September
2010. For the purpose of complete transparency, disclosure information for
the ACCF Board of Trustees, the board of the convening organization of this
document, is available at: http://www.cardiosource.org/ACC/About-ACC/
Leadership/Officers-and-Trustees.aspx. ACCF board members with relevant relationships
with industry to the document may review and comment on the document
but may not vote on approval.
The American College of Cardiology Foundation requests that this document be
cited as follows: Abraham NS, Hlatky MA, Antman EM, Bhatt DL, Bjorkman DJ,
Clark CB, Furberg CD, Johnson DA, Kahi CJ, Laine L, Mahaffey KW, Quigley EM,
Scheiman J, Sperling LS, Tomaselli GF. ACCF/ACG/AHA 2010 expert consensus
document on the concomitant use of proton pump inhibitors and thienopyridines: a
focused update of the ACCF/ACG/AHA 2008 expert consensus document on
reducing the gastrointestinal risks of antiplatelet therapy and NSAID use: a report of
the American College of Cardiology Foundation Task Force on Expert Consensus
Documents. J Am Coll Cardiol. 2010;56:XXX–XX.
This article has been copublished in the American Journal of Gastroenterology and
Circulation.
Copies: This document is available on the World Wide Web sites of the American
College of Cardiology (www.acc.org), the American College of Gastroenterology
(www.acg.gi.org), and the American Heart Association (my.americanheart.org). For
copies of this document, please contact the Elsevier Inc. Reprint Department, fax
212-633-3820, e-mail reprints@elsevier.com.
Permissions: Multiple copies, modification, alteration, enhancement, and/or distribution
of this document are not permitted without the express permission of the
American College of Cardiology Foundation. Please contact Elsevier’s permission
department at healthpermissions@elsevier.com.
Journal of the American College of Cardiology Vol. 56, No. 24, 2010
© 2010 by the American College of Cardiology Foundation, the American College of Gastroenterology,
and the American Heart Association, Inc.
ISSN 0735-1097/$36.00
doi:10.1016/j.jacc.2010.09.010
Published by Elsevier Inc.
ARTICLE IN PRESS
Downloaded from content.onlinejacc.org by on November 9, 2010
TABLE OF CONTENTS
Preamble. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
1.1. Summary of Findings and
Consensus Recommendations. . . . . . . . . . . . . . . . . .XXXX
2. Role of Thienopyridines in CV Disease. . . . . . . . . . . .XXXX
3. Risk of GI Bleeding and Related Mortality
Associated With Clopidogrel Alone
or in Combination. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
4. Strategies to Prevent Thienopyridine-Related
Upper GI Bleeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
4.1. Histamine H2 Receptor Antagonists . . . . . . . . . . .XXXX
4.2. Proton Pump Inhibitors. . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
5. Drug Metabolism: Thienopyridine,
H2RA, and PPI. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
5.1. Thienopyridine Metabolism . . . . . . . . . . . . . . . . . . . . .XXXX
5.2. H2RA Metabolism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
5.3. PPI Metabolism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
6. Hypotheses Regarding the
PPI-Antiplatelet Interaction. . . . . . . . . . . . . . . . . . . . . . . . .XXXX
6.1. Reduced Biological Action of Clopidogrel
Through Competitive Metabolic Effects
of CYP2C19. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
6.2. Reduced Biological Action of Clopidogrel
Related to Genetic Polymorphisms . . . . . . . . . . . .XXXX
7. Evidence-Based Review: PPI and Clopidogrel/
Thienopyridine Pharmacokinetic and
Pharmacodynamic Effect. . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
8. PPI and Clopidogrel/Prasugrel
Clinical Efficacy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
8.1. Do PPIs Decrease Clinical Efficacy of
Clopidogrel or Prasugrel? . . . . . . . . . . . . . . . . . . . . . . .XXXX
8.2. Randomized Clinical Trials . . . . . . . . . . . . . . . . . . . . . .XXXX
8.3. Does the Choice of PPI Matter?. . . . . . . . . . . . . . . .XXXX
8.3.1. Timing of Dosing to
Minimize Interactions . . . . . . . . . . . . . . . . . . . . . . .XXXX
9. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
9.1. The Assessment of Epidemiologic Evidence
Supporting a Significant Clinical Interaction
Between PPIs and Thienopyridines . . . . . . . . . . . .XXXX
9.2. Risk/Benefit Balance: GI Bleed Risk
Versus CV Event Risk . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
9.3. Is H2RA a Reasonable Alternative and in
Which Population?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
9.4. Unanswered Questions and Areas for Future
Research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
Appendix 1. Relevant Author Relationships With
Industry and Other Entities . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
Appendix 2. Relevant Peer Reviewer Relationships
With Industry and Other Entities. . . . . . . . . . . . . . . . . . . . . . .XXXX
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XXXX
Abbreviation List
ACS acute coronary syndromes
ADP adenosine diphosphate
CI confidence interval
CV cardiovascular
GI gastrointestinal
HR hazard ratio
H2RA histamine H2 receptor antagonist
MI myocardial infarction
NNH number-needed-to-harm
NSAID nonsteroidal anti-inflammatory drug
OR odds ratio
PCI percutaneous coronary intervention
PPI proton pump inhibitor
RCT randomized clinical trial
RR relative risk
VASP vasodilator-stimulated phosphoprotein
Preamble
This expert consensus document was developed by the
American College of Cardiology Foundation (ACCF), the
American College of Gastroenterology (ACG), and the
American Heart Association (AHA). Expert consensus
documents inform practitioners, payers, and other interested
parties of the opinion of ACCF and document
cosponsors concerning evolving areas of clinical practice or
medical technologies. Expert consensus documents cover
topics for which the evidence base, experience with technology,
or clinical practice is not considered sufficiently well
developed to be evaluated by the formal ACCF/AHA
Practice Guidelines process. Often, the topic is the subject
of considerable ongoing investigation. Thus, the reader
should view the expert consensus document as the best
attempt of the ACCF and document cosponsors to inform
clinical practice in areas where rigorous evidence may not
yet be available.
To avoid actual, potential, or perceived conflicts of
interest that may arise as a result of industry relationships or
personal interests among the writing committee, all members
of the writing committee, as well as peer reviewers of
the document, are asked to disclose all current health
care-related relationships and those existing 12 months
2 Abraham et al. JACC Vol. 56, No. 24, 2010
Concomitant Use of PPI and Thienopyridines ECD December 7, 2010:000–0
ARTICLE IN PRESS
Downloaded from content.onlinejacc.org by on November 9, 2010
before initiation of the writing effort. The ACCF Task
Force on Clinical Expert Consensus Documents (CECD)
reviews these disclosures to determine which companies
make products (on market or in development) that pertain
to the document under development. Based on this information,
a writing committee is formed to include a majority
of members with no relevant relationships with industry
(RWI), led by a chair with no relevant RWI. Authors with
relevant RWI are not permitted to draft or vote on text or
recommendations pertaining to their RWI. RWI are reviewed
on all conference calls and updated as changes occur.
Author and peer reviewer RWI pertinent to this document
are disclosed in Appendixes 1 and 2, respectively. Additionally,
to ensure complete transparency, authors’ comprehensive
disclosure information—including RWI not pertinent
to this document—is available online. Disclosure information
for the ACCF Task Force on CECD is also available
online at www.cardiosource.org/ACC/About-ACC/
Leadership/Guidelines-and-Documents-Task-Forces.aspx,
as well as the ACCF disclosure policy for document
development at www.cardiosource.org/Science-And-
Quality/Practice-Guidelines-and-Quality-Standards/
Relationships-With-Industry-Policy.aspx.
The work of the writing committee was supported
exclusively by the ACCF without commercial support.
Writing committee members volunteered their time to this
effort. Conference calls of the writing committee were
confidential and attended only by committee members.
1. Introduction
The potential benefits of antiplatelet therapy for atherosclerotic
cardiovascular (CV) disease have been amply demonstrated
over the past 2 decades, especially with regard to the
role of thienopyridine drugs in preventing stent thrombosis.
However, antiplatelet agents increase the risk of bleeding
associated with mucosal breaks in the upper and lower
gastrointestinal (GI) tract. Rational use of thienopyridines is
based on weighing their risks against their benefits. The
magnitude of the risks may vary among patients, based on
their history and clinical characteristics, as may the magnitude
of the benefits.
An earlier Expert Consensus Document, “Reducing the
GI Risks of Antiplatelet and NSAID Use,” recommended
the use of a proton pump inhibitor (PPI) in patients with
risk factors for upper GI bleeding treated with dual antiplatelet
therapy (1). Since its publication, evidence of a
potential adverse drug interaction between PPIs and thienopyridines
has emerged (2). Many recent investigations of
this potential adverse interaction have been performed,
using a variety of research designs. It has been difficult for
practitioners to assimilate this flood of information and to
develop optimal treatment strategies for managing patients
who might benefit from antiplatelet therapy, yet who might
suffer from GI bleeding. The purpose of this document is to
review critically the recent developments in this area, provide
provisional guidance for clinical management, and
highlight areas of future research necessary to address
current knowledge gaps.
1.1. Summary of Findings and
Consensus Recommendations
1. Clopidogrel reduces major CV events compared with
placebo or aspirin.
2. Dual antiplatelet therapy with clopidogrel and aspirin,
compared with aspirin alone, reduces major CV events
in patients with established ischemic heart disease, and
it reduces coronary stent thrombosis but is not routinely
recommended for patients with prior ischemic stroke
because of the risk of bleeding.
3. Clopidogrel alone, aspirin alone, and their combination
are all associated with increased risk of GI bleeding.
4. Patients with prior GI bleeding are at highest risk for
recurrent bleeding on antiplatelet therapy. Other clinical
characteristics that increase the risk of GI bleeding
include advanced age; concurrent use of anticoagulants,
steroids, or nonsteroidal anti-inflammatory drugs
(NSAIDs) including aspirin; and Helicobacter pylori
infection. The risk of GI bleeding increases as the
number of risk factors increases.
5. Use of a PPI or histamine H2 receptor antagonist
(H2RA) reduces the risk of upper GI bleeding compared
with no therapy. PPIs reduce upper GI bleeding
to a greater degree than do H2RAs.
6. PPIs are recommended to reduce GI bleeding among
patients with a history of upper GI bleeding. PPIs are
appropriate in patients with multiple risk factors for GI
bleeding who require antiplatelet therapy.
7. Routine use of either a PPI or an H2RA is not
recommended for patients at lower risk of upper GI
bleeding, who have much less potential to benefit from
prophylactic therapy.
8. Clinical decisions regarding concomitant use of PPIs
and thienopyridines must balance overall risks and
benefits, considering both CV and GI complications.
9. Pharmacokinetic and pharmacodynamic studies, using
platelet assays as surrogate endpoints, suggest that
concomitant use of clopidogrel and a PPI reduces the
antiplatelet effects of clopidogrel. The strongest evidence
for an interaction is between omeprazole and
clopidogrel. It is not established that changes in these
surrogate endpoints translate into clinically meaningful
differences.
10. Observational studies and a single randomized clinical
trial (RCT) have shown inconsistent effects on CV
outcomes of concomitant use of thienopyridines and
PPIs. A clinically important interaction cannot be
excluded, particularly in certain subgroups, such as poor
metabolizers of clopidogrel.
JACC Vol. 56, No. 24, 2010 Abraham et al. 3
December 7, 2010:000–0 Concomitant Use of PPI and Thienopyridines ECD
ARTICLE IN PRESS
Downloaded from content.onlinejacc.org by on November 9, 2010
11. The role of either pharmacogenomic testing or platelet
function testing in managing therapy with thienopyridines
and PPIs has not yet been established.
2. Role of Thienopyridines in CV Disease
Thienopyridine therapy has been evaluated as an alternative
to or in addition to aspirin treatment (“dual antiplatelet
therapy”) to reduce CV events. The absolute risk reduction
from thienopyridines is greater in patients at higher CV
risk, particularly those with acute coronary syndromes
(ACS) or patients who have had a coronary stent implanted.
In patients with ACS without ST-segment elevation,
dual antiplatelet therapy with clopidogrel plus aspirin reduced
the risk of cardiac death, myocardial infarction (MI),
or stroke from 11.4% to 9.3%, compared with aspirin alone,
irrespective of whether patients were revascularized or
treated medically (3) but increased major bleeding from
2.7% to 3.7%. In patients with ST-segment elevation MI
treated with fibrinolytics, the addition of clopidogrel to
aspirin reduced major CV events over 30 days from 10.9%
to 9.1% but increased major bleeding from 1.7% to 1.9%
(4,5).
Dual antiplatelet therapy with aspirin and clopidogrel
reduces stent thrombosis following percutaneous coronary
intervention (PCI) (6). Patients who are implanted with a
bare-metal stent are recommended to receive at least 1
month of clopidogrel, and patients receiving a drug-eluting
stent are recommended to receive dual therapy for at least 12
months. In patients with atrial fibrillation who are unable to
take vitamin-K antagonists, adding clopidogrel to aspirin
reduced the rate of major vascular events (7.6% to 6.8%) and
stroke (3.3% to 2.4%) compared with aspirin alone but with
a greater risk of bleeding—2.0% per year (7).
In patients with established atherosclerotic CV disease,
clopidogrel alone reduced (5.8% to 5.3%) the combined risk
of major CV events, ischemic stroke, MI, and vascular death
compared with aspirin alone (8) and led to less GI bleeding
(2.7% to 2.0%). Clopidogrel is recommended as an alternative
agent for patients with CV disease unable to take
aspirin (9–12).
In the primary prevention setting, dual antiplatelet therapy
with clopidogrel plus aspirin did not significantly reduce
major CV events compared with aspirin alone (6.8% versus
7.3%) but increased severe bleeding (1.3% to 1.7%) (13).
Patients with recent ischemic stroke or transient ischemic
attack treated with clopidogrel plus aspirin had an insignificant
reduction in major CV events (16.7% to 15.7%)
compared with aspirin alone and experienced more lifethreatening
hemorrhages (1.3% to 2.6%) (14).
Prasugrel is a new thienopyridine derivative with a rapid
onset and consistent inhibition of platelet aggregation. In
patients with ACS and planned PCI, prasugrel reduced
major CV events from 12.1% to 9.9% compared with
clopidogrel but increased major bleeding from 1.8% to 2.4%
and fatal bleeding from 0.1% to 0.4% (15).
Ticagrelor, a novel, reversible, direct-acting P2Y12 receptor
blocker (not yet approved for use in the United States)
reduced the primary endpoint of vascular death, MI, or
stroke from 11.7% to 9.8% compared with clopidogrel, with
no significant difference in major bleeding (11.6% versus
11.2%) but with an increased risk of noncoronary artery
bypass graft major bleeding (3.8% to 4.5%) (16).
For patients with ischemic stroke or transient ischemic
attack, antiplatelet therapy with aspirin, clopidogrel, or the
combination of dipyridamole and aspirin is recommended
to prevent recurrent stroke, but the combination of clopidogrel
and aspirin is not recommended (17), and prasugrel
is contraindicated (15).
3. Risk of GI Bleeding and Related Mortality
Associated With Clopidogrel Alone
or in Combination
GI bleeding among patients receiving antiplatelet therapy
can develop from many different lesions and anatomic sites.
Upper GI bleeding may be due to esophagitis (18) or peptic
ulcer disease related to H. pylori infection, or aspirin or other
NSAIDs (19). These mucosal breaks are aggravated by the
antiplatelet effects of thienopyridines, promoting bleeding.
Bleeding from other GI sites is also exacerbated by antiplatelet
therapy (20–27).
Several risk factors for GI bleeding in the setting of
antiplatelet therapy have been reported consistently. A
history of bleeding or other complications of peptic ulcer
disease is the strongest risk factor for subsequent upper GI
bleeding (28). Advanced age also significantly increases the
absolute risk of upper GI bleeding. Use of anticoagulants,
steroids, or NSAIDs has also been shown to be consistent
predictors for GI bleeding, as has H. pylori infection
(29–35). The relative risk (RR) of GI bleeding increases
with the number of adverse risk factors present in an
individual patient (36).
The risk of GI bleeding associated with thienopyridines
has been assessed in several case-control studies (Online
Table 1) and in RCTs with prospectively assessed GI
bleeding safety endpoints (Online Table 2). In head-tohead
randomized trials of aspirin and clopidogrel, the risk of
GI bleeding was higher in patients treated with aspirin
(Online Table 2), although the absolute risk difference was
small.
Dual antiplatelet therapy with clopidogrel and aspirin
increased the risk of GI bleeding by 2- to 3-fold compared
with aspirin alone in randomized trials (Online Table 2),
but the absolute risk increase was in the range of 0.6% to
2.0%. Two RCTs (3,7) provide specific data on GI bleeding
risk associated with dual antiplatelet therapy, demonstrating
an RR of 1.78 (95% confidence interval [CI]: 1.25 to 2.54;
number needed to harm [NNH] of 130) and 1.96 (95% CI:
4 Abraham et al. JACC Vol. 56, No. 24, 2010
Concomitant Use of PPI and Thienopyridines ECD December 7, 2010:000–0
ARTICLE IN PRESS
Downloaded from content.onlinejacc.org by on November 9, 2010
1.46 to 2.63; NNH of 167). There are fewer data on the risk
of GI bleeding in routine practice among patients who are
less selected and not as closely monitored as patients in
clinical trials. In a cohort of Tennessee Medicaid patients
treated with clopidogrel, the rate of upper GI bleeding was
1.2% per year (36).
There are few data on the mortality attributable to GI
bleeding in patients on clopidogrel alone or on dual antiplatelet
therapy. In studies of varying duration and design,
the case fatality rates for GI bleeding associated with dual
antiplatelet therapy have been low (0% to 0.3%) (3,29–31).
Nevertheless, the RR for death from a GI bleed has been
estimated at 2.5 (37), and GI bleeding appears to be a
significant predictor of death, even after adjustment for CV
morbidity, age, sex, diabetes, PCI status, and concomitant
therapy (37,38).
4. Strategies to Prevent
Thienopyridine-Related Upper GI Bleeding
Thienopyridines do not cause ulcers or erosions of the
digestive tract (39), but their antiplatelet effects may promote
bleeding at the site of preexisting lesions caused by the
use of aspirin or NSAIDs, or infection with H. pylori (40).
Upper GI bleeding in the setting of thienopyridine use may
be reduced by suppressing gastric acid production, thereby
promoting healing of peptic ulcers and mucosal erosions, as
well as by stabilizing thrombi (41). Acid production can be
suppressed either by H2RAs or by PPIs; the efficacy of each
has been examined to prevent GI bleeding related to
antiplatelet use.
4.1. Histamine H2 Receptor Antagonists
The use of H2RAs can suppress gastric acid production by
37% to 68% over 24 hours (42,43), and standard doses have
a modest protective effect in patients taking aspirin. In a
randomized trial of 404 patients with peptic ulcers or
esophagitis who were taking aspirin, fewer gastroduodenal
ulcers developed over 12 weeks among patients assigned to
famotidine (3.8%) than to placebo (23.5%; p 0.0002) (18).
In another study, however, H2RAs did not significantly
protect clopidogrel users (RR: 0.83; 95% CI: 0.20 to 3.51)
(44). No randomized trials have directly compared PPIs
with H2RAs in patients with CV disease on antiplatelet
therapy. However, observational data suggest PPIs may be
more effective than H2RAs in preventing upper GI bleeding.
In a cohort of 987 patients who were prescribed aspirin
and clopidogrel, PPI use led to a greater reduction in upper
GI bleeding (odds ratio [OR]: 0.04; 95% CI: 0.002 to 0.21)
than H2RA use (OR: 0.43; 95% CI: 0.18 to 0.91) (30).
4.2. Proton Pump Inhibitors
PPIs reduce gastric acid secretion for up to 36 hours (45).
Observational data suggest that PPIs reduce the risk of GI
bleeding in patients on antiplatelet therapy. In 1 cohort
study, the baseline clopidogrel-related gastroduodenal
bleeding risk of 1.2% per year was reduced by 50% in
patients prescribed a PPI (36). In this same study, PPI use
reduced the absolute risk of GI bleeding by 2.8% per year
among patients with 3 risk factors for GI bleeding. In a
large case-control study comparing 2,779 patients with
endoscopically confirmed upper GI hemorrhage with 5,532
controls, concomitant use of a PPI and a thienopyridine was
associated with less upper GI bleeding (RR: 0.19; 95% CI:
0.07 to 0.49) than thienopyridine use alone (44). Smaller
cohort studies confirm similar risk reduction with concurrent
PPI prescription (31). In the results of a recent
randomized trial (46), patients with CV disease taking
enteric-coated aspirin who were randomized to receive
clopidogrel plus omeprazole had fewer GI events (i.e., a
composite outcome of overt or occult bleeding, symptomatic
gastroduodenal ulcer or erosion) than patients randomized
to receive clopidogrel alone (hazard ratio [HR]: 0.34; 95%
CI: 0.18 to 0.63).
5. Drug Metabolism: Thienopyridine,
H2RA, and PPI
5.1. Thienopyridine Metabolism
Clopidogrel is a pro-drug converted in vivo to an active
metabolite that irreversibly binds to the platelet adenosine
diphosphate (ADP) P2Y12 receptor, thereby inhibiting
platelet aggregation. The bioavailability of the active metabolite
is determined by intestinal absorption, which may
be influenced by an ABCB1 polymorphism, and by metabolism
through the cytochrome P-450 pathway (47). Clopidogrel
is activated in a 2-step process (Figure 1A) mediated
by oxidative biotransformation in the liver, in which
CYP2C19 and CYP3A have particularly important roles
(Figure 1A) (48,49). The parent compound clopidogrel, and
to a lesser extent 2-oxo-clopidogrel, are both substrates and
inhibitors of CYP1A2, CYP2B6, and CYP2C19 (50).
Clopidogrel and 2-oxo-clopidogrel are extensively hydrolyzed
to inactive metabolites, potentially magnifying the
effects of CYP2C19 inhibitors and polymorphisms (51).
However, redundant pathways (Figure 1A) for activation of
clopidogrel may mitigate the effect of inhibitors and reduced
function polymorphisms of CYP450 isoenzymes in vitro
(49,52).
Prasugrel is also a pro-drug that requires biotransformation
to active metabolites by cytochrome P-450 enzymes,
including CYP3A isoforms, CYP2B6, CYP2C9, and
CYP2C19 (Figure 1A). Prasugrel is hydrolyzed to a thiolactone
derivative in the intestine and then oxidized to its active
metabolite in both the intestine and the liver (Figure 1A)
(51,53). Reduced-function CYP2C19 alleles are not believed
to have a clinically meaningful effect in prasugreltreated
patients (54).
JACC Vol. 56, No. 24, 2010 Abraham et al. 5
December 7, 2010:000–0 Concomitant Use of PPI and Thienopyridines ECD
ARTICLE IN PRESS
Downloaded from content.onlinejacc.org by on November 9, 2010
Ticagrelor (AZD6140) is an orally active cyclopentyltriazolopyrimidine
adenosine triphosphate analog that reversibly
inhibits P2Y12 platelet receptors (Figure 1B). Ticagrelor,
which is not yet approved in the United States, is an
active compound and is metabolized by CYP3A4 to an
active metabolite (55,56). Ticagrelor and its active metabolite
are both metabolized and glucuronidated in the liver
before elimination in the urine. Genetic variations in CYP
isoenzymes do not appear to affect metabolism of ticagrelor.
Other frequently used CV medications are also metabolized
by the CYP450 system (51,52) and may interact with
thienopyridine metabolism. Of note are statins, which are
metabolized by the CYP450 system (51,52), and aspirin,
which induces CYP2C19 (57).
5.2. H2RA Metabolism
The H2RAs currently available in the United States (cimetidine,
ranitidine, famotidine, and nizatidine) vary in their
ability to inhibit gastric acid secretion. Hepatic metabolism
is the dominant elimination pathway for orally administered
cimetidine (60%), ranitidine (73%), and famotidine (50% to
80%) but not nizatidine (22%) (58). Cimetidine may interact
with drugs metabolized via the cytochrome P-450
pathway, as it inhibits CYP1A2, 2C9, 2C19, 2D6, 2E1, and
3A4 (59–61). Although cimetidine might decrease the
biotransformation of clopidogrel by competitive inhibition
of CYP2C19, there have been no controlled studies of this
hypothesis. Ranitidine interacts weakly with cytochrome
P-450 (58,62,63), and famotidine and nizatidine do not
bind to the cytochrome P-450 system and, therefore, have
low potential to interact with clopidogrel (58,62).
5.3. PPI Metabolism
All PPIs used in the United States (omeprazole, esomeprazole,
pantoprazole, rabeprazole, lansoprazole, and dexlansoprazole)
are weak bases converted to their active forms in
the acidic environment of active gastric parietal cells (64).
PPIs are metabolized by the hepatic cytochrome P-450
Figure 1. Thienopyridine Metabolism
(A) Clopidogrel and prasugrel; (B) ticagrelor. ATP indicates adenosine triphosphate; CYP, cytochrome P-450; and hCE1 and 2, human carboxylesterases 1 and 2.
6 Abraham et al. JACC Vol. 56, No. 24, 2010
Concomitant Use of PPI and Thienopyridines ECD December 7, 2010:000–0
ARTICLE IN PRESS
Downloaded from content.onlinejacc.org by on November 9, 2010
system, predominantly CYP2C19, and, to a lesser extent,
CYP3A4 (65). The studies assessing the degree to which
different PPIs interact with CYP2C19 have yielded inconsistent
results, so no definitive conclusions can be drawn
comparing the pharmacokinetics and potential for drug
interaction of the various PPIs with clopidogrel and
prasugrel.
6. Hypotheses Regarding the
PPI-Antiplatelet Interaction
6.1. Reduced Biological Action of Clopidogrel
Through Competitive Metabolic Effects of CYP2C19
Concomitant use of PPIs may competitively inhibit activation
of clopidogrel by CYP2C19, thereby attenuating its
antiplatelet effect. Coadministration of other CYP2C19-
inhibiting drugs may further reduce the efficacy of clopidogrel
and inhibition of platelet aggregation (66). The
reported interaction of clopidogrel and PPIs is consistent
with a set of clinical pharmacokinetic findings referred to as
high-risk pharmacokinetics (66). The risk of drug inefficacy
is greater when drug concentrations depend on variable
activity of a single metabolic pathway.
6.2. Reduced Biological Action of Clopidogrel
Related to Genetic Polymorphisms
The potential for impaired antiplatelet activity is supported
by data on the effect of natural variations in CYP2C19
activity, based on genetic polymorphisms. The CYP2C19*2,
CYP2C19*3, and CYP2C19*4 alleles decrease active metabolite
production compared with the most common CYP2C19
genotype. Individuals who are heterozygous for loss-offunction
alleles are “intermediate metabolizers,” and those who
are homozygous are “poor metabolizers.” CYP2C19 polymorphisms
have been associated with reduced platelet inhibition
and an increased rate of recurrent CV events (53,67,68).
Reduced platelet inhibition may be overcome with higher
clopidogrel doses (69), but any increased CV efficacy from
higher-dose treatment must be weighed against an increased
risk of GI bleeding (70).
The best characterized and most common loss-offunction
polymorphism is the CYP2C19*2 allele (53),
which is carried by 51% to 55% of Asians, 33% to 40% of
African Americans, 24% to 30% of Caucasians, and 18% of
Mexican Americans (53,71–75). The antiplatelet effect of
clopidogrel varies directly with the number of loss-offunction
alleles; 2 copies are associated with a 65% reduction
in clopidogrel antiplatelet efficacy and 1 copy with a 47%
reduction (71–75). The genetic variation in CYP2C19 is
associated with up to a 50% greater risk of adverse clinical
outcomes, including CV death, MI, or stroke, and a 3-fold
increased risk of stent thrombosis in patients receiving
clopidogrel (53,72). However, the CYP2C19*2 variant appears
to account for only 12% of variation in platelet
aggregability in response to ADP; and other factors, such as
diabetes, obesity, and acute ischemia (76), likely contribute
much more to variability in platelet response (72,73,77).
7. Evidence-Based Review:
PPI and Clopidogrel/Thienopyridine
Pharmacokinetic and Pharmacodynamic Effect
Platelet function tests serve as surrogate markers for the
clinical effectiveness of antiplatelet drugs. The standard
platelet function test is aggregometry, which measures
ADP-stimulated platelet aggregation in whole blood or
platelet-rich plasma. A more recent test quantifies phosphorylation
of vasodilator-stimulated phosphoprotein
(VASP) in whole blood and appears to be a more specific
measure of clopidogrel-mediated inhibition of platelet aggregation.
The newest test, the Verify Now P2Y12 assay, is
similar to VASP. It has not been established that changes in
these surrogate endpoints translate into clinically meaningful
differences.
Among 162 healthy subjects, carriers of at least 1
reduced-function CYP2C19 allele had significantly less
inhibition of platelet aggregation on standard aggregometry
in response to clopidogrel than did noncarriers (53). The
ultrarapid metabolizer genotypes had the greatest platelet
inhibition from clopidogrel, and the poor metabolizer genotypes
had the least platelet inhibition.
The influence of omeprazole on the antiplatelet effects of
clopidogrel was assessed in a double-blind trial (78) of 124
patients after coronary stenting treated with aspirin and
clopidogrel. Patients randomized to omeprazole for 7 days
had significantly less platelet inhibition, as measured by the
VASP method, than patients randomized to placebo. In
another study of 104 patients given a higher maintenance
dose of 150 mg clopidogrel after coronary stenting (79),
patients randomized to omeprazole had significantly less
platelet inhibition on the VASP assay than patients randomized
to pantoprazole, with 44% clopidogrel nonresponders
in the omeprazole group compared with 23% in
the pantoprazole group (p 0.04). In the PRINCIPLE–
TIMI 44 (Prasugrel in Comparison to Clopidogrel for
Inhibition of Platelet Activation and Aggregation–
Thrombolysis In Myocardial Infarction 44) trial, patients
undergoing PCI taking a PPI had significantly less platelet
inhibition with clopidogrel than those not on a PPI,
whereas patients taking prasugrel as well as a PPI had a
trend toward reduced-platelet inhibition (80).
In randomized trials that used ex vivo platelet assays as
surrogate clinical endpoints, patients treated with omeprazole
demonstrated impaired clopidogrel response (78,79),
even when a high antiplatelet dose was used. Studies of
other PPIs have not demonstrated this effect (79,81), but
these studies were conducted in different populations using
different study designs. Few direct head-to-head comparison
studies have been reported. The ongoing SPICE
JACC Vol. 56, No. 24, 2010 Abraham et al. 7
December 7, 2010:000–0 Concomitant Use of PPI and Thienopyridines ECD
ARTICLE IN PRESS
Downloaded from content.onlinejacc.org by on November 9, 2010
(Evaluation of the Influence of Statins and Proton Pump
Inhibitors on Clopidogrel Antiplatelet Effects) trial
(NCT00930670) will directly compare the effects of commonly
prescribed PPIs (i.e., omeprazole, pantoprazole, esomeprazole)
and a H2RA (ranitidine) on ex vivo platelet
aggregation among 320 post-PCI patients who require dual
antiplatelet therapy. Secondary outcomes include assessment
of clopidogrel resistance, prevalence of CYP2C19*2
polymorphism and its effect on PPI and antiplatelet activity,
all-cause mortality, MI, revascularization, stroke, and GI
bleeding at 1 year (82).
8. PPI and Clopidogrel/Prasugrel
Clinical Efficacy
8.1. Do PPIs Decrease Clinical Efficacy of
Clopidogrel or Prasugrel?
Observational studies of different populations, sizes, and
degree of methodologic rigor have examined whether patients
prescribed a PPI plus clopidogrel have an increased
risk of CV events compared with patients prescribed clopidogrel
alone (Online Table 3). The results are mixed: several
studies have shown small but significant associations between
PPI use and CV events, but others show no
significant association. The magnitude of the treatment
effect in positive studies has been modest, with risk ratios
2.0. Whether differences in study results are because of
differences in confounding factors between study groups
cannot be determined. In observational studies, PPIs may
be selectively prescribed to higher-risk patients, potentially
biasing the estimated CV risk (36). Small, yet
significant, differences in common, clinically important
events would, however, represent an important public
health issue.
The effect of PPIs on clinical efficacy has been evaluated
retrospectively in nonrandomized cohorts within randomized
trials. In a study of 13,608 patients randomized to
either clopidogrel or prasugrel after PCI, use of PPI did not
affect the outcome of a composite of CV death, MI, or
stroke, either among clopidogrel-assigned patients (adjusted
HR: 0.94; 95% CI: 0.80 to 1.11) or among the prasugrelassigned
patients (HR: 1.00; 95% CI: 0.84 to 1.20) (80). In
this study, there was no difference among the PPIs used,
including omeprazole (n 1,675), lansoprazole (n 441),
esomeprazole (n 613), and pantoprazole (n 1,844). The
results were similar among those with a reduced-function
CYP2C19 allele. In the CREDO (Clopidogrel for Reduction
of Events During Observation) trial, PPI use was
associated with an increased rate of CV events whether or
not the patient was treated with clopidogrel (83). The
evidence from these studies and observational comparisons
is inconclusive regarding the clinical effects of concomitant
use of a PPI and a thienopyridine.
8.2. Randomized Clinical Trials
Only 1 RCT has examined the potential interaction between
clopidogrel and PPIs with CV events as the outcome.
In a double-blind, placebo-controlled trial (46), 3,761
patients with either ACS or PCI were randomized to a
fixed-dose combination of clopidogrel and omeprazole
(75/20 mg) or clopidogrel alone. All patients received
aspirin. The data from this trial revealed no significant
difference in a composite CV endpoint (MI, stroke, coronary
artery bypass graft, PCI, CV death) for patients on the
fixed-dose combination compared with clopidogrel alone
(HR: 0.99; 95% CI: 0.68 to 1.44), but fewer GI adverse
events (HR: 0.34; 95% CI: 0.18 to 0.63). However, the
study was halted short of its planned enrollment and
duration; and the number of CV events was low (55 versus
54 CV events). Consequently, the confidence limits for CV
events are broad and do not exclude a clinically important
increase in risk of up to 44%.
8.3. Does the Choice of PPI Matter?
Pharmacokinetic studies in vitro have suggested that all
PPIs inhibit CYP2C19 to varying degrees, but the relative
magnitude of inhibition varies by specific PPI and laboratory
assay used. Pharmacodynamic studies using ADPstimulated
platelet aggregation in patients treated with
clopidogrel suggest a variable inhibitory effect of different
PPIs (80,84,85), but few head-to-head comparison studies
have been performed.
In the combined analysis of 2 trials of clopidogrel and
prasugrel, the rate of CV death, MI, or stroke was similar
for all PPIs and no different than the rate in patients not
taking a PPI (80). A nested case-control study of patients
receiving clopidogrel after MI suggested pantoprazole may
increase the risk of rehospitalization for MI or PCI compared
with other PPIs (86). However, a retrospective cohort
study of 20,596 patients showed no effect of any PPI on the
frequency of CV events among patients taking clopidogrel,
with similar HRs for esomeprazole, lansoprazole, omeprazole,
pantoprazole, and rabeprazole (36). Other observational
studies of patients taking clopidogrel have suggested
that the risk of CV events is similar for all PPIs (45,87,88).
Thus, although pharmacokinetic and pharmacodynamic
data suggest varying inhibition by different PPIs of the
enzyme systems necessary to convert clopidogrel to its active
form, there is no good evidence that these differences on
surrogate markers translate into meaningful differences in
clinical outcomes. No prospective trials directly compare the
clinical events of different PPIs in patients treated with
clopidogrel.
8.3.1. Timing of Dosing to Minimize Interactions
Because the plasma half-lives of both clopidogrel and all
available PPIs are less than 2 hours, interactions between
these drugs might be minimized by separating the timing of
drug administration, even among poor CYP2C19 metabolizers
(45). In a crossover study examining 72 healthy
8 Abraham et al. JACC Vol. 56, No. 24, 2010
Concomitant Use of PPI and Thienopyridines ECD December 7, 2010:000–0
ARTICLE IN PRESS
Downloaded from content.onlinejacc.org by on November 9, 2010
subjects who were administered standard-dose clopidogrel
(300 mg followed by 75 mg daily) and a supratherapeutic
dose of omeprazole (80 mg daily), mean inhibition of
platelet aggregation was greater when the drugs were given
12 hours apart (89). Further studies will be required to
evaluate this hypothesis, using appropriate drug doses and
meaningful clinical endpoints. Until data from such studies
are available, there is no solid evidence to recommend that
the dosing of PPIs be altered.
9. Conclusions
9.1. The Assessment of Epidemiologic Evidence
Supporting a Significant Clinical Interaction
Between PPIs and Thienopyridines
When assessing a possible causal link between an exposure
and an outcome, it is recommended to consider: 1) the
strength of the association, 2) consistency of the association
across different samples, 3) existence of a biologically plausible
mechanism of action, and 4) supportive experimental
evidence (90). In applying these principles to the concomitant
use of PPIs and thienopyridines, we draw the following
conclusions:
1. The magnitude of association in positive observational
studies reviewed is small to moderate (HR or OR: 2),
but associations of this magnitude in nonrandomized
observational studies may be due to residual differences
in patient characteristics between study groups. Large,
well-controlled randomized trials are necessary to
assess the validity of small-to-moderate magnitude
associations. The only available randomized trial
showed no significant association of omeprazole with
CV events, but the confidence limits on this null
finding include the possibility of up to a 44% relative
increase in CV risk.
2. A significant association between PPI use and increased
CV events has been inconsistently demonstrated in
observational studies, with the majority of studies showing
no association. In addition, available studies markedly
vary in methodologic rigor.
3. Although clinical studies with CV events as endpoints
are not definitive, the proposed mechanism is biologically
plausible, given that a) clopidogrel users with
reduced-function genetic polymorphisms in CYP2C19
metabolism have increased rates of CV events; and b) in
vitro testing suggests that PPIs may inhibit CYP2C19
metabolism.
4. Experimental pharmacodynamic data consistently indicate
that omeprazole diminishes the effect of clopidogrel
on platelets. Other pharmacodynamic studies have failed
to demonstrate a significant effect of other PPIs on
clopidogrel. In the absence of large-scale, randomized,
experimental studies that directly compare PPIs with
different pharmacokinetic properties, the evidence remains
weak for diminished antiplatelet activity associated
with PPIs and thienopyridine coprescription. The
ongoing SPICE trial may provide additional answers and
address issues regarding the clinical relevance of such
interactions.
9.2. Risk/Benefit Balance: GI Bleed Risk
Versus CV Event Risk
All prescription drugs have favorable and unfavorable effects,
and treatment decisions must be based on whether the
potential for benefit outweighs the potential for harm. The
CV benefits of antiplatelet drugs are overwhelmingly documented
for patients who have ACS and patients who
undergo PCI. It is also well demonstrated that antiplatelet
drugs increase the risk of GI bleeding. The magnitude
of these benefits and risks in individual patients varies
depending on their characteristics (36). The challenge for
healthcare providers is to determine the risk/benefit
balance for individual patients or subsets of the target
population.
PPIs are coprescribed with antiplatelet drugs for 1 reason—
to reduce the increased risk of GI complications
caused by antiplatelet drugs. The need for GI protection
increases with the number of risk factors for severe bleeding.
Prior upper GI bleeding is the strongest and most consistent
risk factor for GI bleeding on antiplatelet therapy. Patients
with ACS and prior upper GI bleeding are at substantial
CV risk, so dual antiplatelet therapy with concomitant use
of a PPI may provide the optimal balance of risk and
benefit. Among stable patients undergoing coronary revascularization,
a history of GI bleeding should inform the
choice of revascularization method; if a coronary stent is
selected to treat such patients, the risk/benefit tradeoff may
favor concomitant use of dual antiplatelet therapy and a
PPI.
Advanced age; concomitant use of warfarin, steroids, or
NSAIDs; or H. pylori infection all raise the risk of GI
bleeding with antiplatelet therapy. The risk reduction with
PPIs is substantial in patients with risk factors for GI
bleeding and may outweigh any potential reduction in the
CV efficacy of antiplatelet treatment because of a drug– drug
interaction. Patients without these risk factors for GI
bleeding receive little if any absolute risk reduction from a
PPI, and the risk/benefit balance would seem to favor use of
antiplatelet therapy without concomitant PPI. The reduction
of GI symptoms by PPIs (i.e., treatment of dyspepsia)
may also prevent patients from discontinuing their antiplatelet
treatment. The discontinuation of antiplatelet therapy
in patients with GI bleeding may increase the risk of
CV events (91).
9.3. Are H2RAs a Reasonable Alternative
and in Which Population?
H2RAs are effective compared with placebo in decreasing
the risk of gastric and duodenal ulcers (92) caused by
NSAIDs and antiplatelet therapy (18), but not as effective as
PPIs (93,94). PPIs are also more effective than H2RAs for
JACC Vol. 56, No. 24, 2010 Abraham et al. 9
December 7, 2010:000–0 Concomitant Use of PPI and Thienopyridines ECD
ARTICLE IN PRESS
Downloaded from content.onlinejacc.org by on November 9, 2010
preventing ulcers in patients using high doses of NSAIDs
(95) and are effective in decreasing GI bleeding in patients
prescribed aspirin or thienopyridines (36,96,97). Available
data suggest PPIs are superior to H2RAs, but H2RAs may
be a reasonable alternative in patients at lower risk for GI
bleeding, and in those who do not require PPI for refractory
gastroesophageal reflux disease. Cimetidine can competitively
inhibit CYP2C19, so other H2RAs might be a better
choice in patients treated with clopidogrel.
9.4. Unanswered Questions and Areas for
Future Research
Many gaps in knowledge exist regarding GI bleeding
among patients prescribed thienopyridines. The pathophysiology
of GI hemorrhage associated with thienopyridines is
not fully understood and should be further elucidated.
Better data are needed on the incidence of GI bleeding
among patients taking antiplatelet therapy, particularly in
relation to clinical factors that may alter the risk of bleeding.
The tradeoffs between bleeding risk and cardiovascular
benefits of antiplatelet therapy deserve further study. Clinical
trials of strategies to reduce the risk of GI bleeding
among patients with CV disease on antiplatelet therapy,
particularly using the commonly prescribed PPIs and highdose
H2RAs, would provide direct evidence on the comparative
effectiveness of alternative management strategies.
There is considerable variation among patients in response
to antiplatelet therapy, so the potential role of
laboratory testing in individualization of therapy should be
a high priority for research. Either pharmacogenomic testing
for CYP2C19 variants or platelet function testing might
be used to tailor therapy by guiding the choice of drug
(thienopyridines, PPIs, H2RAs), the choice of drug dose, or
both. Although the concept of individually tailored therapy
is rational and attractive, empirical evidence for this approach
is sparse. Clinical studies and randomized trials
comparing guided therapy with usual care are needed, as are
trials comparing different approaches to guided therapy
(e.g., pharmacogenomic profiling versus platelet function
testing). Studies that compare different management options
for patients with specific test results would also be
useful: For example, what are the effects on clinical outcomes
of using a higher dose of clopidogrel among patients
who are either “poor metabolizers” on a genetic test or who
have relatively little platelet inhibition on a functional assay?
Finally, we need to evaluate the effect on clinical outcomes
of dosing schedules that minimize simultaneous exposure to
high levels of a PPI and a thienopyridine.
10 Abraham et al. JACC Vol. 56, No. 24, 2010
Concomitant Use of PPI and Thienopyridines ECD December 7, 2010:000–0
ARTICLE IN PRESS
Downloaded from content.onlinejacc.org by on November 9, 2010
