Risk of adverse cardiovascular outcomes and all-cause mortality associated with concomitant use of clopidogrel and proton pump inhibitors in elderly patients
All rights reserved: reproduction in whole or part not permitted
Original articleRisk of adverse cardiovascular outcomes andall-cause mortality associated with concomitantuse of clopidogrel and proton pump
of concomitant use of clopidogrel and PPIs in a national sample of elderly Medicare
A nested case–control design was employed. A cohort of Medicare beneficiaries who initiated clopidogrel
days between clopidogrel fills between July 1, 2006 and December 31,
national sample of Medicare claims data. Within this cohort, cases
Address for correspondence: Yi Yang MD PhD,
(beneficiaries who experienced any major cardiovascular event [MCE] [acute myocardial infarction,
stroke, coronary artery bypass graft, or percutaneous coronary intervention] or all-cause mortality) and
University of Mississippi, University, MS 38677, USA. 20
controls (beneficiaries who did not experience any MCE or all-cause mortality) were identified from inpatient
Tel.: þ1 (662) 915-1062; Fax: þ1 (662) 915-5102;
claims. Cases and controls were matched on age and the time to first clopidogrel fill.
Conditional logistic regression was performed on the matched sample to evaluate the associationbetween concomitant use of clopidogrel and PPIs and adverse health outcomes (MCEs and all-cause
Clopidogrel – Drug–drug interaction – Elderly – Sale or
A total of 43,159 clopidogrel users were identified. Among them, 15,415 (35.7%) received clopidogrel and
Curr Med Res Opin Downloaded from informahealthcare.com by 207.68.251.244 on 02/27/13
a PPI concomitantly at any time during the study period, 3502 (8.1%) experienced a MCE, 7306 (17.1%)
died, and a total of 9908 (22.8%) experienced the primary composite outcome (any MCE or all-cause
disp mortality) during follow-up. The odds ratio (OR) for the primary composite outcome was 1.26 (95%
confidence interval [CI]: 1.18–1.35). Secondary analyses indicated that elderly patients using clopidogreland a PPI concomitantly were more likely to experience all-cause mortality (OR: 1.40; 95% CI: 1.29–1.53)as compared to those receiving clopidogrel only, but not MCEs (OR: 1.06; 95% CI: 0.95–1.18).
Concomitant use of clopidogrel and PPIs was associated with a slightly increased risk of all-cause mortalitybut not MCEs.
Clopidogrel belongs to the class of anti-platelet agents and is indicated for pre-vention of arterial thromboembolism in patients with peripheral arterial diseasesor patients with acute coronary syndrome (ACS) who are treated medically or
Concomitant use of clopidogrel and PPIs in the elderly
through coronary revascularization procedures1,2. With
younger than 50 years of age28 and physicians often pre-
global sales of $9.1 billion (US dollars) in 2009, clopido-
scribe PPIs to these patients to provide gastrointestinal
grel is the second best selling drug worldwide3.
protection29 because PPIs have been found to be the
Clopidogrel is highly effective in reducing athero-
most effective agent in preventing gastrointestinal bleed-
thrombotic and ischemic events; however, its use is often
ing related to anti-platelet therapy30. Research on the con-
associated with gastrointestinal complications including
comitant use of clopidogrel and PPIs focusing on the
ulceration and bleeding4. Several professional organiza-
elderly population has been limited. This lack of informa-
tions such as the American Heart Association and the
tion is concerning because the effect of concomitant use of
American College of Gastroenterology have recom-
clopidogrel and PPIs in the elderly cannot be simply extra-
mended that a proton pump inhibitor (PPI) should be
polated from studies involving younger people. A few
used with anti-platelet therapy consisting of clopidogrel
observational studies have evaluated the effect of clopido-
alone or clopidogrel and aspirin for gastrointestinal protec-
grel–PPI interaction in the elderly population and results
tion5. However, evidence from mechanistic studies has
from these studies are likely to be inadequate to inform
indicated that the anti-platelet effect of clopidogrel is
evidence-based prescribing because these studies only
reduced in the presence of PPIs because PPIs inhibit
included patients within restricted geographic areas31,32.
CYP2C19, an enzyme responsible for the conversion of
In a randomized clinical study, Bhatt et al. examined the
clopidogrel to its active metabolite6–8. Moreover, several
incidence of adverse gastrointestinal events and adverse
observational studies have shown that concomitant use
cardiovascular events among elderly patients on dual clo-
of clopidogrel and PPIs increases the risk of adverse car-
pidogrel and aspirin therapy receiving either omeprazole or
diovascular events9–15. In response to growing concerns
placebo, and they did not find apparent cardiovascular
over combined use of clopidogrel and PPIs, the US Food
and Drug Administration (FDA) issued a public advisory
While this randomized controlled study provides useful
in 2009 advising physicians to re-evaluate the need of PPI
use in patients receiving clopidogrel and subsequently
among elderly patients, only the interaction between clo-
a ‘black box’ warning was added to all clopidogrel
pidogrel and omeprazole and not other PPIs was evaluated
(Plavix*) labels in 2011 advising physicians to avoid pre-
in the study. To provide further insights into the real-world
scribing clopidogrel and PPIs, namely omeprazole and
association between concomitant use of clopidogrel and
esomeprazole, simultaneously16. Similarly, the European
PPIs and adverse outcomes in the elderly population, we
Medicines Agency issued a public statement in 2009 dis-
conducted a nested case–control study involving a
couraging the concomitant use of clopidogrel and PPIs
national sample of elderly patients using the 5% sample
unless absolutely necessary17. However, no consensus has
of Medicare administrative claims data.
been reached regarding the effect of concurrent use ofclopidogrel and PPIs, as results from two randomized con-trolled trials and several observational studies have shown
no association between concomitant use of clopidogrel andPPIs and adverse cardiovascular events5,8,18–24.
Most previous research on concomitant use of clopido-
The study used a nested case–control design for the
grel and PPIs has focused on general patient populations.
analysis of Medicare beneficiaries with age
Curr Med Res Opin Downloaded from informahealthcare.com by 207.68.251.244 on 02/27/13
A better understanding of the effects of the potential drug–
who initiated and continued to receive clopidogrel therapy
drug interaction between clopidogrel and PPIs focusing
with or without a PPI between July 1, 2006 and December
on the elderly is important for several reasons. First, the
31, 2008. The study was approved by the University of
effect of concomitant use of clopidogrel and PPIs is likely
Mississippi Institutional Review Board and the data were
to be different in older adults, as older adults are more
accessed only after execution of a data use agreement with
likely to experience cardiovascular events25 and they are
the Centers for Medicare and Medicaid Services (CMS).
also more susceptible to clopidogrel–PPI interactions26. Elderly patients might be more likely to be influenced bythe clopidogrel–PPI interaction possibly due to their
reduced liver function, which reduces metabolism ofclopidogrel to its active metabolite27. When a PPI is
The 2006–2008 5% national sample of Medicare admin-
co-administered with clopidogrel in the elderly, clopido-
istrative claims database was used for this study. Records
grel’s anti-platelet effect could be further reduced. Second,
for healthcare services offered to Medicare beneficiaries in
elderly patients on anti-platelet drugs are more likely to
various settings such as inpatient, outpatient, and skilled
have gastrointestinal bleeding as compared to those
nursing facilities were available through the Medicareclaims database. The database also included records for
*Plavix is a registered trade name of Sanofi-Aventis, NJ, USA.
prescription drugs dispensed under Medicare Part D.
2 Concomitant use of clopidogrel and PPIs in the elderly Mahabaleshwarkar et al.
Beneficiary demographic and enrollment information was
identify PCI. We identified AMI based on the approach
available through the Medicare Beneficiary Summary file.
proposed by Kiyota et al. (2004)35, which has been found to
All claims were linked through an encrypted unique
have a high positive predictive value. For the identifica-
common beneficiary identification number.
tion of stroke, we used the algorithm suggested by Rekeret al. (2001)36. Both AMI and stroke were identified usingMedicare inpatient claims. All-cause mortality was deter-
mined using information from the Medicare Beneficiary
The cohort used for this study consisted of Medicare ben-
Summary file. For each case, we selected one control
from the cohort who did not experience a MCE or all-
2006; (2) had continuous Medicare part A coverage and at
cause mortality; cases and their corresponding controls
least 1 month of Part B coverage from January 1, 2006 to
were matched on age (Æ5 years) and time to cohort
December 31, 2008 or until death, whichever occurred
entry (Æ7 days). Time to cohort entry was determined as
first; and (3) had initiated clopidogrel therapy and did
the time between the start of the study period, July 1, 2006,
not have any gap of 30 days or more between clopidogrel
and the date of the first clopidogrel fill. The index date
fills between July 1, 2006 and December 31, 2008, since
for each control was set as the index date of the corres-
previous literature suggests that a gap of 30 days or
ponding case to which it was matched (i.e., the date on
more between prescription fills should be considered as
which a MCE or all-cause mortality occurred).
non-persistence with therapy33. Clopidogrel fills wereidentified from the Medicare Part D drug event file using
the National Drug Codes for clopidogrel. Only incidentusers of clopidogrel, that is, beneficiaries who had no
Information about use of clopidogrel and PPIs for each
clopidogrel claims between January 1 and June 30, 2006
patient was obtained from the Medicare Part D drug
(i.e., the baseline period), were included in the study.
event file. The PPIs considered in this study included ome-
Beneficiaries enrolled in Medicare health maintenance
prazole, esomeprazole, pantoprazole, lansoprazole, and
organizations for any part of the study period were also
rabeprazole. Concomitant use of clopidogrel and PPIs in
excluded since no medical claims were available for
cases and controls was determined based on whether or not
the beneficiary possessed the two drugs simultaneouslybetween July 1, 2006 and the index date. Possession
of the drugs was determined on a daily basis using the
prescription fill date and the days’ supply field for each
In this study, our primary end point was a composite
prescription. Early fills were taken into account by carrying
outcome of any major cardiovascular event (MCE) or
forward the days of overlap between prescription fills.
all-cause mortality. MCEs considered in this study were
Concomitant use of clopidogrel and PPIs was considered
acute myocardial infarction (AMI), stroke, receiving cor-
as a dichotomous variable. Beneficiaries who possessed
onary artery bypass grafting (CABG) surgery, or receiving
clopidogrel and a PPI together for one or more days
percutaneous coronary intervention (PCI). For patients
between July 1, 2006 and the index date were considered
experiencing more than one adverse event, the first occur-
as concomitant users of clopidogrel and PPIs, whereas
rence of any of these events was deemed as the index
those who did not possess the two drugs simultaneously
Curr Med Res Opin Downloaded from informahealthcare.com by 207.68.251.244 on 02/27/13
event. Within the cohort of clopidogrel users, we defined
were considered as clopidogrel only users.
cases as patients who experienced an index event betweenthe date of the first prescription claim for clopidogrel and
December 31, 2008. The date that a beneficiary experi-enced the index event was designated as the index date
Covariates included in the study were baseline demo-
for cases. Cases were identified using the International
graphics (gender, and race), presence of end stage renal dis-
ease (ESRD), dual eligibility status, presence of important
Modification (ICD9-CM) diagnosis and procedure codes
cardiovascular comorbidities (angina, chronic obstructive
and the CMS Common Procedure Coding System
pulmonary disease [COPD], congestive heart failure,
(HCPCS) codes from Medicare inpatient and outpatient
diabetes, hypertension, hyperlipidemia, and peripheral
claims data (Appendix 1). CABG was identified based on
arterial disease), previous use of dialysis, and history of
the primary and secondary procedure codes associated with
the inpatient records. Since PCI may be performed as an
bleeding, duodenal or gastric ulcer, gastroesophageal
outpatient procedure with a patient being discharged on
reflux disease, and erosive esophagitis), AMI, stroke,
the same day34, we used primary and secondary procedure
CABG, and PCI. Race was divided into three categories:
codes associated with inpatient and outpatient claims to
white, black, and others. Presence of ESRD and dual
Concomitant use of clopidogrel and PPIs in the elderly
eligibility were both considered as dichotomous variables.
conditions (such as cancer) also contribute to all-cause
Presence of cardiovascular comorbidities, history of gastro-
mortality. Most of these conditions have been taken into
intestinal disorders, previous use of dialysis, history of
consideration in the calculation of CCI and previous
AMI, stroke, CABG, and PCI were determined based
research has shown that CCI is a significant predictor of
on the inpatient and outpatient claims at baseline
all-cause mortality38. Finally, we evaluated the association
(January 1 – June 30, 2006). In addition, we adjusted for
between concomitant use of clopidogrel and individual
the use of classes of medications which may affect the anti-
types of PPIs and the occurrence of the primary composite
platelet effect of clopidogrel including other CYP2C19
outcome of any MCE or all-cause mortality. All analyses
inhibitors (chloramphenicol, cimetidine, felbamate, flu-
were performed using Statistical Analysis System (SAS)
oxetine, fluvoxamine, ketoconazole, moedafinil, and
version 9.2 (SAS Institute Inc., Cary, NC, USA). The
oxcarbazepine), CYP3A4 inhibitors (amiodarone, diltia-
greedy algorithm was used for matching cases and controls.
zem, verapamil, voriconazole, fluconazole, nicardipine,
and nifedipine), and sulfonylureas during the study
STRATA statement was used for conditional logistic
period (July 1, 2006 – December 31, 2008)37. We also
adjusted for the use of drugs which have the potential toreduce the likelihood of adverse cardiovascular eventssuch as angiotensin-converting-enzyme inhibitors, angio-
tensin-receptor antagonists, acetylsalicylic acid, -adre-nergic antagonists, calcium-channel antagonists, digoxin,
We identified 43,159 Medicare beneficiaries as incident
spironolactone, statins, and diuretics during the study
clopidogrel users. Of these beneficiaries, 15,415 (35.7%)
used clopidogrel and a PPI concomitantly at any timeduring the study period; 3502 (8.1%) experienced aMCE, 7306 (17.1%) died, and a total of 9908 (23.0%)
experienced an adverse outcome (a MCE or all-cause mor-tality) during follow-up. The majority of the beneficiaries
In the bivariate analyses, the chi-square test was used for
were female (61.5%) and white (84.0%). The mean age of
the comparison of categorical variables whereas Student’s
the cohort was 76.8 years. Detailed patient demographic
t-test was used for the comparison of continuous variables
and clinical characteristics are presented in Table 1.
between cases and controls. Frequencies and percentages
Baseline demographic and comorbidity characteristics
were reported for categorical variables, whereas means
of the matched cases and controls are also presented in
and standard deviations were reported for continuous vari-
Table 1. For example, cases were more likely than controls
ables. In the multivariable analysis, conditional logistic
to be males (39.5% vs. 33.8%, p50.001) and blacks (8.9%
regression was performed on matched cases and controls
vs. 7.9%, p50.001). A greater percentage of cases had
to determine the association between concomitant clopi-
diabetes (32.6% vs. 29.4%, p50.001), COPD (17.5% vs.
dogrel and PPI use and the occurrence of any MCE or all-
12.8%, p50.001), and congestive heart failure (20.7% vs.
cause mortality. The covariates listed earlier were adjusted
12.5%, p50.001) than controls. A greater percentage of
for in the conditional logistic regression model. Odds ratios
controls had hypertension (68.8% vs. 61.8%, p50.001),
(OR) and 95% confidence intervals (CIs) were computed.
hyperlipidemia (52.7% vs. 40.7%, p50.001), erosive
In addition, we conducted a set of secondary analyses
esophagitis (3.6% vs. 2.7%, p ¼ 0.002), and gastroesopha-
Curr Med Res Opin Downloaded from informahealthcare.com by 207.68.251.244 on 02/27/13
to test the robustness of our primary findings. Firstly,
geal reflux disease (13.2% vs. 11.0%, p50.001) than cases.
we evaluated the association between concomitant use of
In terms of medication use, a greater percentage of cases
clopidogrel and PPIs and the occurrence of each individual
were users of CYP3A4 inhibitors (5.2% vs. 4.5%,
adverse outcome (AMI, stroke, CABG, PCI, and all-cause
p ¼ 0.037), angiotensin-converting enzyme inhibitors
mortality). Secondly, we computed the odds of experien-
(51.7% vs. 46.7%, p50.001), -adrenergic antagonists
cing any MCE associated with concomitant use of clopi-
(74.0% vs. 68.6%, p50.001), and sulfonylureas (18.3%
dogrel and PPIs. For patients having more than one MCE,
vs. 15.3%, p50.001) than controls. Controls were more
the first occurrence of any of these events was defined as
likely to be users of angiotensin receptor antagonists
the index event. Thirdly, while determining the associ-
(22.5% vs. 19.9%, p50.001) and statins (67.3% vs.
ation between concomitant use of clopidogrel and PPIs
60.3%, p50.001) as compared to cases.
and all-cause mortality, instead of using each individual
From our primary analysis with conditional logistic
cardiovascular comorbidity, the Charlson comorbidity
regression performed on matched cases and controls,
index (CCI, D’Hoore adaptation)38 and cardiovascular
after extensive multivariable adjustment, we found that
comorbidities not included in the CCI calculation –
the estimated odds of experiencing the primary composite
namely angina, hyperlipidemia, and hypertension – were
outcome (any MCE or all-cause mortality) were 26%
used as covariates, because various non-cardiovascular
higher for patients with concomitant use of clopidogrel
4 Concomitant use of clopidogrel and PPIs in the elderly Mahabaleshwarkar et al.
Table 1. Baseline characteristics of incident clopidogrel users and matched cases and controls.
Medication use between July 1, 2006 and December 31, 2008Other CYP2C19 inhibitors
aFor continuous variables, p values are based on t-tests for differences between means; for categorical variables, p values are based on Pearson’s chi-square testof association. bOther races include Asian, Hispanic, and North American natives. ESRD ¼ end stage renal disease; COPD ¼ chronic obstructive pulmonary disease; AMI ¼ acute myocardial infarction; CABG ¼ coronary artery bypass grafting;
Curr Med Res Opin Downloaded from informahealthcare.com by 207.68.251.244 on 02/27/13
PCI ¼ percutaneous coronary intervention.
and a PPI as compared to patients who were taking clopi-
obtained by including individual cardiovascular comorbid-
dogrel only (OR: 1.26; 95% CI: 1.18–1.34, p50.001).
ities as covariates (OR: 1.36; 95% CI: 1.25–1.47,
In the secondary analyses, we found a significant effect
p50.001). The results evaluating the effects of individual
of concomitant use of clopidogrel and a PPI on all-cause
types of PPIs were similar to the results of our primary
mortality (OR: 1.40; 95% CI: 1.29–1.53, p50.001),
whereas the effects on individual cardiovascular eventswere not statistically significant (Table 2). The associationbetween concomitant use of clopidogrel and PPIs and
occurrence of any MCE was not statistically significant(OR: 1.06; 95% CI: 0.95–1.18, p ¼ 0.305). In terms of
This study examined the risk of adverse outcomes asso-
association between concomitant use of clopidogrel and
ciated with concomitant use of clopidogrel and PPIs in a
PPIs and all-cause mortality, the results on including
national sample of elderly Medicare beneficiaries. To our
CCI and cardiovascular comorbidities not included in
knowledge, this is the first study evaluating the effect of
the CCI calculation as covariates were similar to those
clopidogrel–PPI interaction in the elderly population
Concomitant use of clopidogrel and PPIs in the elderly
Table 2. Association between concomitant use of clopidogrel and proton
Medicare beneficiaries using clopidogrel and a PPI con-
pump inhibitors and adverse health outcomes.
comitantly as compared to beneficiaries using clopidogrelonly. Results from our secondary analyses indicate that
concomitant use of clopidogrel and PPIs is associated
with significantly increased risk for all-cause mortality,
but not for MCEs. The significant association between
concomitant clopidogrel and PPI use and occurrence of
all-cause mortality held even after controlling for CCI in
the secondary analysis. These results suggest that all-cause
mortality was the major contributor to our primary finding
of statistically significant association between concomi-
tant use of clopidogrel and PPI and adverse health out-
comes. These results are somewhat consistent with the
aObtained from conditional logistic regressions performed on cases and
findings from two previous studies focusing on the elderly
controls matched on age (Æ5 years) and time to cohort entry (Æ7 days). bAdjusted for all patient characteristics (covariates) in Table 1 except for
population. Rassen et al. (2009) evaluated the clopidogrel–
age, time to cohort entry, and Charlson comorbidity index.
PPI interaction in low-income patients enrolled in three
PPI ¼ proton pump inhibitor; MCE ¼ major cardiovascular event (AMI,
health insurance programs in British Columbia, New
CABG ¼ coronary artery bypass grafting; PCI ¼ percutaneous coronary
Jersey, and Pennsylvania32. They studied patients aged
!65 years who were hospitalized for ACS or PCI between2001 and 2005. They found that concomitant use of clo-pidogrel and PPIs was associated with an increased but not
Table 3. Association between concomitant use of clopidogrel and
statistically significant risk of myocardial infarction or
individual types of proton pump inhibitors and major adverse cardiovascularevents or all-cause mortalitya.
death (OR: 1.22; 95% CI: 0.99–1.51). The associationsbetween concurrent clopidogrel and PPI use and occur-
rence of individual outcomes of death (OR: 1.20; 95%
CI: 0.84–1.70) and revascularization (OR: 0.97; 95% CI:
0.79–1.21) were also found to be non-significant in this
study. In a Canadian study, Juurlink et al. (2009) studied
elderly AMI patients who were prescribed clopidogrel
within 3 days of discharge from a local hospital. Theyfound a significant association between readmission for
aAdjusted for all patient characteristics (covariates) in Table 1 except for
AMI and current use of a PPI (most recent prescription
age, time to cohort entry, and Charlson comorbidity index. bObtained from conditional logistic regressions performed on cases and
fill for a PPI within 30 days before readmission for AMI or
controls matched on age (Æ5 years) and time to cohort entry (Æ7 days).
death) (OR: 1.27; 95% CI: 1.03–1.57). However, a signifi-
Major cardiovascular event ¼ AMI, stroke, CABG, or PCI.
cant association between readmission for AMI and PPI use
AMI ¼ acute myocardial infarction; CABG ¼ coronary artery bypass grafting;PCI ¼ percutaneous coronary intervention.
was not observed by the authors in previous users of clopi-dogrel, who had their most recent prescription fill for a PPI31–90 days before readmission for AMI or death, and in
using a national sample. Elderly patients on clopidogrel
remote users of clopidogrel, who had their most recent
Curr Med Res Opin Downloaded from informahealthcare.com by 207.68.251.244 on 02/27/13
therapy have an increased tendency to gastrointestinal
prescription fill for a PPI 91–180 days before readmission
bleeding and hence need effective medications such as
for AMI or death31. As far as the effect of the clopidogrel–
PPIs for gastrointestinal protection. On the other hand,
PPI interaction in the general population is concerned,
they may be more prone to be affected by a potential clo-
some observational studies have found an increased risk
pidogrel–PPI interaction. Although a previous randomized
of cardiovascular events9–15, whereas others have found
controlled study has examined the potential interaction
no statistically significant effect19–24. Since these studies
between clopidogrel and omeprazole in an elderly popula-
consisted of patient populations that included younger
tion18, this study adds to the literature by determining
individuals, direct comparison of findings from our study
the real-world effect of clopidogrel–PPI interaction in
with these studies was not possible.
Regarding the effect of concomitant use of clopidogrel
Through a nested case–control design, we studied the
and individual types of PPIs on adverse health outcomes,
association between concomitant use of clopidogrel and
the results were statistically significant across all types of
PPIs and the risk of experiencing a MCE or all-cause mor-
PPIs. The significant interaction between clopidogrel and
tality using the 5% national sample of Medicare benefici-
pantoprazole was a little unexpected, given that previous
aries. We found that the risk of experiencing this
mechanistic studies and observational studies evaluating
composite outcome was 26% higher among elderly
the potency of PPIs in CYP2C19 inhibition showed that
6 Concomitant use of clopidogrel and PPIs in the elderly Mahabaleshwarkar et al.
pantoprazole had the least inhibitory potency39–44.
gastrointestinal disorders or a physician’s impression that
Significant interaction between clopidogrel and pantopra-
a given patient is at an increased risk for gastrointestinal
zole was also observed in studies by Kreutz et al.11 and
problems. Furthermore, in our analyses, we could not
Stockl et al.45, suggesting that a different mechanism
account for unmeasured factors such as smoking, blood
may be associated with the interaction between clopido-
pressure level, serum cholesterol level, and family history
grel and PPIs. Results concerning the clopidogrel–panto-
of cardiovascular disease, which are likely to affect cardio-
prazole interaction in studies focusing on elderly adults
vascular outcomes. We used Medicare beneficiaries’ pre-
have been inconsistent. While Rassen et al.32 observed
scription drug filling data to identify medication uses.
no difference in rate ratios for adverse outcomes in panto-
Though the beneficiaries possessed the medications, we
prazole and other PPI users, Juurlink et al.31 found that
were not able to verify whether they actually took them.
current pantoprazole users were less likely to experience
Nonetheless, previous research has shown that filling a
adverse outcomes compared to other PPI users. Future
prescription is usually consistent with taking the drug46.
efforts are needed to further explore the mechanism under-
Finally, causes of death were not available from our data
lying the interaction between clopidogrel and PPIs.
therefore all-cause mortality, and not cardiovascular mor-
Our study has a number of strengths. To our knowledge,
tality, was used as an outcome measure in our study.
this is the only study evaluating the effect of concomitantuse of clopidogrel and PPIs in older adults using a nationalsample. Second, we have included all clopidogrel users inthe study unlike other studies which have concentrated on
a specific set of users such as individuals hospitalized for
We found a slightly increased risk of experiencing any
ACS or PCI32 or AMI31. We believe this broad-based
MCE or all-cause mortality in elderly adults using clopido-
patient inclusion makes findings from our study more gen-
grel and PPIs concomitantly. These results were consistent
eralizable. Furthermore, we have taken utmost care to
across all PPIs. The associations between concomitant clo-
eliminate many sources of confounding in our study.
pidogrel and PPI use and occurrence of MCEs, considered
For example, we have adjusted for the presence of major
individually or as a group, were not statistically significant,
cardiovascular comorbidities such as hypertension, dia-
while the odds of all-cause mortality were higher in elderly
betes, and congestive heart failure, which usually lead to
concomitant users of clopidogrel and PPIs. Since the
adverse cardiovascular events. We also adjusted for the use
reasons of death were unidentifiable from our data,
of medications which may potentially affect clopidogrel’s
considering the fact that all-cause mortality can occur
anti-platelet effect. Additionally, we included CCI as a
due to non-cardiovascular reasons, the results for the asso-
covariate in our analysis to determine the effect of clopi-
ciation between concomitant clopidogrel and PPI use and
dogrel–PPI interaction on all-cause mortality.
all-cause mortality might have been biased away from
Several limitations of our study need to be addressed.
the null in spite of controlling for CCI in the secondary
Due to limitations of the data source, we could not identify
analyses. Therefore, findings from this study do not present
from our cohort, beneficiaries who were poor metabolizers
conclusive evidence to suggest avoiding PPIs in elderly
of clopidogrel due to genetic polymorphisms of the cyto-
patients on anti-platelet therapy with clopidogrel.
chrome P450 enzyme CYP2C19. Poor clopidogrel metab-olizers may be more likely to experience adverse
Conservatively, other gastrointestinal protecting agents
cardiovascular outcomes or mortality. Using Medicare
such as histamine-2-receptor antagonists could be
Curr Med Res Opin Downloaded from informahealthcare.com by 207.68.251.244 on 02/27/13
administrative claims data, we were also unable to identify
over-the-counter use of PPI medications (omeprazole andlansoprazole). It is possible that clopidogrel only userswould be more likely to have purchased over-the-counter
PPIs than those who have received prescription PPIs. Similarly, we were unable to account for the over-
the-counter use of aspirin, which is generally used with
This study was sponsored by the Centers for Medicare and
clopidogrel to treat peripheral arterial diseases or ACS
because aspirin is often available over the counter. In add-
Declaration of financial/other relationships
ition, although we have tried to minimize confounding by
R.K.M., Y.Y., M.V.D., J.P.B., M.W.S., B.F.B., and K.D.N. have
indication through controlling for history of gastrointes-
disclosed that they have no significant relationships with or
tinal disorders (including gastrointestinal bleeding, duo-
financial interests in any commercial companies related to this
denal or gastric ulcer, gastroesophageal reflux disease,
and erosive esophagitis) in our analyses, there may be
CMRO peer reviewers may have received honoraria for
other sources of confounding by indication which are
their review work. The peer reviewers on this manuscript have
not captured by our data such as uncoded diagnoses of
disclosed that they have no relevant financial relationships.
Concomitant use of clopidogrel and PPIs in the elderly
This study was conducted as part of a project titled ‘Using
Medicare/Medicaid Claims Data to Support Medication
Outcomes and Pharmacovigilance Research’, which was sup-
Public statement on possible interaction between clopidogrel and proton
ported by grant award 1COCMS330731/01 from the Office of
pump inhibitors. London: European Medicines Agency, 2009. Available at:
Research, Development and Information, CMS.
Previous presentation: This work was presented at the
Bhatt DL, Cryer BL, Contant CF, et al. Clopidogrel with or without omeprazole
International Society for Pharmacoeconomics and Outcomes
in coronary artery disease. N Engl J Med 2010;363:1909-17
Research 17th Annual International Meeting, 2–6 June 2012,
Ray WA, Murray KT, Griffin MR, et al. Outcomes with concurrent use of
clopidogrel and proton-pump inhibitors. Ann Intern Med 2010;152:337-45
Schmidt M, Johansen MB, Robertson DJ, et al. Concomitant use of clopido-grel and proton pump inhibitors is not associated with major adverse cardio-vascular events following coronary stent implantation. Aliment Pharmacol
Hsiao FY, Mullins CD, Wen YW, et al. Relationship between cardiovascular
Aschenbrenner DS. Drug watch. Am J Nurs 2006;106:29-31
outcomes and proton pump inhibitor use in patients receiving dual antiplatelet
Joseph JJ, Pillai A, Bramley D. Clopidogrel in orthopaedic patients: a review of
therapy after acute coronary syndrome. Pharmacoepidemiol Drug Saf
current practice in Scotland. Thromb J 2007;5:6
Top 15 Global Products, 2009, Total Audited Markets. IMS Heath Reports.
Charlot M, Ahlehoff O, Norgaard ML, et al. Proton-pump inhibitors are asso-
New Jersey, US: IMS Health, 2009. Available at:
ciated with increased cardiovascular risk independent of clopidogrel use: a
nationwide cohort study. Ann Intern Med 2010;153:378-86
Tentzeris I, Jarai R, Farhan S, et al. Impact of concomitant treatment with
proton pump inhibitors and clopidogrel on clinical outcome in patients after
Gurbel PA, Tantry US, Kereiakes DJ. Interaction between clopidogrel and
coronary stent implantation. Thromb Haemost 2010;104:1211-18
proton-pump inhibitors and management strategies in patients with cardio-
Rossini R, Capodanno D, Musumeci G, et al. Safety of clopidogrel and proton
vascular diseases. Drug Healthc Patient Saf 2010;2:233-40
pump inhibitors in patients undergoing drug-eluting stent implantation. Coron
Harjai KJ, Shenoy C, Orshaw P, et al. Clinical outcomes in patients with the
concomitant use of clopidogrel and proton pump inhibitors after percutaneous
Cardiovascular disease risk factors. Geneva: World Heart Federation.
coronary intervention. Circ Cardiovasc Interv 2011;4:162-70
Gilard M, Arnaud B, Cornily JC, et al. Influence of omeprazole on the anti-
platelet action of clopidogrel associated with aspirin: the randomized, double-
Delafuente JC. Understanding and preventing drug interactions in elderly
blind OCLA (Omeprazole CLopidogrel Aspirin) study. J Am Coll Cardiol2008;51:256-60
patients. Crit Rev Oncol Hematol 2003;48:133-43
Li XQ, Andersson TB, Ahlstro¨m M, et al. Comparison of inhibitory effects of
Wynne HA, James OFW. The ageing liver. Age Ageing 1990;19:1-3
Nikolsky E, Stone GW, Kirtane AJ, et al. Gastrointestinal bleeding in patients
the proton pump-inhibiting drugs omeprazole, esomeprazole, lansoprazole,pantoprazole, and rabeprazole on human cytochrome P450 activities.
with acute coronary syndromes: incidence, predictors, and clinical implica-
tions: analysis from the ACUITY (Acute Catheterization and Urgent Intervention
O’Donoghue ML, Braunwald E, Antman EM, et al. Pharmacodynamic
Triage Strategy) trial. J Am Coll Cardiol 2009;54:1293-302
effect and clinical efficacy of clopidogrel and prasugrel with or without
Abrahamsen B, Eiken P, Eastell R. Proton pump inhibitor use and the anti-
a proton-pump inhibitor: an analysis of two randomized trials. Lancet
fracture efficacy of alendronate. Arch Intern Med 2011;171:998-1004
Ng FH, Tunggal P, Chu WM. Esomeprazole compared with famotidine in
Gupta E, Bansal D, Sotos J, et al. Risk of adverse clinical outcomes with
the prevention of upper gastrointestinal bleeding in patients with acute
concomitant use of clopidogrel and proton pump inhibitors following percu-
coronary syndrome or myocardial infarction. Am J Gastroenterol 2012;107:
taneous coronary intervention. Dig Dis Sci 2010;55:1964-8
Ho PM, Maddox TM, Wing L, et al. Risk of adverse outcomes associated with
Juurlink DN, Gomes T, Ko DT, et al. A population-based study of the drug
Curr Med Res Opin Downloaded from informahealthcare.com by 207.68.251.244 on 02/27/13
concomitant use of clopidogrel and proton pump inhibitors following acute
interaction between proton pump inhibitors and clopidogrel. CMAJ 2009;
Kreutz RP, Stanek EJ, Aubert R, et al. Impact of proton pump inhibitors on the
Rassen JA, Choudhry NK, Avorn J, Schneeweiss S. Cardiovascular outcomes
effectiveness of clopidogrel after coronary stent placement: the Clopidogrel
and mortality in patients using clopidogrel with proton pump inhibitors after
Medco Outcomes Study. Pharmacotherapy 2010;30:787-96
percutaneous coronary intervention or acute coronary syndrome. Circulation
Tsai Y, Wen Y, Huang W, et al. Cardiovascular and gastrointestinal events of
three antiplatelet therapies: clopidogrel, clopidogrel plus proton-pump inhibi-
Sørensen R, Gislason GH, Fosbøl EL, et al. Initiation and persistence with
tors, and aspirin plus proton-pump inhibitors in patients with previous gastro-
clopidogrel treatment after acute myocardial infarction: a nationwide study. Br
intestinal bleeding. J Gastroenterol 2011;46:39-45
Van Boxel OS, Van Oijen MGH, Hagenaars MP, et al. Cardiovascular and
Rao SV, Kaltenbach LA, Weintraub WS, et al. Prevalence and outcomes of
gastrointestinal outcomes in clopidogrel users on proton pump inhibitors:
same-day discharge after elective percutaneous coronary intervention among
results of a large Dutch cohort study. Am J Gastroenterol 2010;105:2430-6
Yasuda H, Yamada M, Sawada S, et al. Upper gastrointestinal bleeding in
Kiyota Y, Schneeweiss S, Glynn RJ, et al. The accuracy of Medicare claims-
patients receiving dual antiplatelet therapy after coronary stenting. Intern Med
based diagnosis of acute myocardial infarction: estimating positive predictive
value based on review of hospital records. Am Heart J 2004;148:99-104
Bhurke SM, Martin BC, Li C, et al. Effect of the clopidogrel-proton pump
Reker DM, Hamilton BB, Duncan PW, et al. Stroke: Who’s counting what? J
inhibitor drug interaction on adverse cardiovascular events in patients with
acute coronary syndrome. Pharmacotherapy 2012;32:809-18
Harmsze AM, van Werkum JW, Moral F, et al. Sulfonylureas and on-clopido-
Clopidogrel (marketed as Plavix) and Omeprazole (marketed as Prilosec) –
grel platelet reactivity in type 2 diabetes mellitus patients. Platelets 2011;22:
drug interaction. FDA Safety Information. Washingon, DC: US Food and Drug
8 Concomitant use of clopidogrel and PPIs in the elderly Mahabaleshwarkar et al.
D’Hoore W, Sicotte C, Tilquin C. Risk adjustment in outcome assessment: the
Siller-Matula JM, Spiel AO, Lang IM, et al. Effects of pantopraole and esome-
Charlson comorbidity index. Method Inform Med 1993;32:382-7
prazole on platelet inhibition by clopidogrel. Am Heart J 2009;157:148.e1-5
Li XQ, Andersson TB, Ahlstrom M, et al. Comparison of inhibitory effects of the
Angiolillo DJ, Gibson CM, Cheng S, et al. Differential effects of omeprazole
proton pump-inhibiting drugs omeprazole, esomeprazole, lansoprazole, pan-
and pantoprazole on the pharmacodynamics and pharmacokinetics of clopi-
toprazole, and rabeprazole on human cytochrome P450 activities. Drug
dogrel in healthy subjects: randomized, placebo-controlled, crossover com-
parison studies. Clin Pharmacol Ther 2011;89:65-74
Cuisset T, Frere C, Quilici J, et al. Comparison of omeprazole and pantopra-
Stockl KM, Le L, Zakharyan A, et al. Risk of rehospitalization for patients
zole influence on a high 150-mg clopidogrel maintenance dose. The PACA
using clopidogrel with a proton pump inhibitor. Arch Intern Med 2010;170:
(Proton Pump Inhibitors and clopidogrel Association) prospective randomized
Steiner JF, Prochazka AV. The assessment of refill compliance using phar-
Sibbing D, Morath T, Stegherr J, et al. Impact of proton pump inhibitors on the
macy records: methods, validity, and applications. J Clin Epidemiol 1997;50:
antiplatelet effects of clopidogrel. Thromb Haemost 2009;101:714-19
Neubauer H, Engelhardt A, Kruger JC, et al. Pantoprazole does not influence
Tran M, Tafreshi J, Pai RG. Combination of clopidogrel and proton pump
the antiplatelet effect of clopiodgrel: a whole blood aggregometry study after
inhibitors: implications for clinicians. J Cardiovasc Pharmacol Ther
coronary stenting. J Cardiovasc Pharmacol 2010;56:91-7
Appendix 1. ICD9-CM and HCPCS codes used to identify cardiovascular outcomes.
primary diagnosis code of 430.xx–434.xx, 436.xx, 433.01,
433.11, 433.21, 433.31, 433.81, or 433.91 or primary diag-nosis code of V57.xx and any secondary diagnosis code of342.xx, 430.xx, 431.xx, 432.xx, 433.xx, 434.xx, 435.xx,436.xx, 437.xx or 438.xx or primary diagnosis code of 433.xxor 435.xx and any secondary diagnosis code of 342.xx, 430.xx,431.xx, 432.xx, 434.xx, or 436.xxa
33510–33514, 33516, 33533–33536, 33517–33519 33521–33523,
36.01, 36.02, 36.05, 00.66, 36.06, 36.07b
92982, 92984, 92995, 92996, 92980, 92981, 92973–92975, 92978,
aICD9-CM diagnosis codes. bICD9-CM procedure codes. CABG ¼ coronary artery bypass graft; PCI ¼ percutaneous coronary intervention; AMI ¼ acute myocardial infarction; ICD9-CM ¼ International Classification forDiseases version 9 Clinical Modification; HCPCS ¼ Centers for Medicare and Medicaid Services Common Procedure Coding System.
Curr Med Res Opin Downloaded from informahealthcare.com by 207.68.251.244 on 02/27/13
Concomitant use of clopidogrel and PPIs in the elderly
DHS DRUG FORMULARY FOR THE HWLA INITIATIVE - BY DRUG NAME July 2012 Bolded and italicized denote Patient Assistance Program (PAP) availability Medication Therapeutic Class Additional Information $4 Program PAP Availability Allergies, Cold and Flu Guaifenesin/ Dextromethorphan 100mg-10mg/5mL Oral Solution Anticonvulsant Divalproex delayed release 125mg sprinkle ca
RESUMEN-COMENTARIO DE LA CUESTIÓN 94 A Tomás de Aquino se le conoce como al autor de una síntesis entre aristotelismo y cristianismo. Bien, una síntesis no es lo mismo que una coincidencia plena: ambas éticas tienen carácter teleólogico (toda la argumentación se construye sobre el fin del ser humano), ambas son éticas materiales, ambas sitúan a la felicidad como el fin suprem