Journal reference: van Boxel OS, van Oijen MGH, Hagenaars MP, et al. Cardiovascular and gastrointestinal outcomes in clopidogrel users on proton pump inhibitors: results of a large Dutch cohort study.  Am J Gastroenterol 2010; 105: 2430–6 ^[1]

Link: http://dx.doi.org/10.1038/ajg.2010.334

Published: 24 August 2010

Evidence cookie says...

Concurrent proton pump inhibitors (PPIs) may increase the risk of cardiovascular outcomes in patients taking clopidogrel. The results of this specific study are likely invalid due to methodological biases (see below for details). There is substantial uncertainty in the evidence base generally. GPs should be guided by a precautionary principle: it is reasonable to prescribe PPIs to those with clear indications new and ongoing therapy should be assessed from the perspective that PPIs might increase the risk of harm

Article details:

Study design:

retrospective cohort study

Study aim:

to investigate the association between the co-administration of a proton pump inhibitor (PPI) and clopidogrel (Plavix, Iscover), and the occurrence of cardiovascular (CV) and gastrointestinal (GI) events in a large cohort in the Netherlands

Study conclusion:

New clopidogrel users on PPIs are at an increased risk of cardiovascular and GI complications compared with those who are not using a PPI. The inferior cardiovascular profile of clopidogrel users on PPIs and the occurrence of channeling bias may be important factors underlying this observation ^[1].

Critical appraisal:

Methodology (PICO):

Participants: who was studied?

• 18,139 patients
  • data were collected from anonymised computerized databases of two Dutch health insurance companies
  • total pool of 4 million Dutch inhabitants (~ 25 % of the total population)
  • 1 January 2006 to 31 December 2007.
  • ≥ 18 years
  • ≥ 1 year of valid history available in the database
  • ≥ 1 prescription for clopidogrel
• PPI use:
  • ≥ 80% overlap between PPI and clopidogrel use
  • or, concurrent use of both medications within 7 days before or at the time of a possible event
• exclusions:
  • patients with documented use of clopidogrel in the 180 days before the index clopidogrel prescription date

Concurrent PPI and clopidogrel user group:

• 5,734 patients
• 58.5% male
• mean age: 68.6 years

Clopidogrel users without PPI group:

• 12,405 used clopidogrel without a PPI
• 66.9% male
• mean age: 66.1 years

Notable differences between groups:

• patients on concurrent PPIs were:
  • significantly older
  • had more comorbidities (including heart failure, history of unstable angina pectoris, chronic obstructive pulmonary disease and renal failure)
  • used more co-medications

Intervention: what was the exposure?

• clopidogrel, and
• PPI:
  • omeprazole (Losec, Acimax)
  • pantoprazole (Somac)
  • lansoprazole (Zoton)
  • rabeprazole (Pariet)
  • esomeprazole (Nexium)

Comparator: what was the control/alternative?

• clopidogrel

Outcomes: what was measured?

Primary outcome:

• composite of:
  • myocardial infarction
  • unstable angina pectoris
  • stroke
  • all-cause mortality
• during clopidogrel use or within 7 days of cessation

Secondary outcome:

• individual components of the primary outcome
• occurrence of peptic ulcer disease (PUD)
• during clopidogrel use or within 7 days of cessation

Are the trial results valid?

Internal validity: Wikipedia

Was the defined representative sample of patients assembled at a common point in the course of their disease?

Unclear.

• the number of subjects which formed the basis of the study was substantial; 4 million Dutch inhabitants
• however, as these were insured individuals (data through insurance company database), it is unclear whether these patients are representative of the overall population

Were the groups similar at the start of the trial?

No.

• there were substantial differences between the two groups
• the concurrent PPI and clopidogrel group were older and had more medical co-morbidities
• this is a serious threat to the internal validity of the study

Was patient follow up sufficiently long and complete?

N/A

Were outcome criteria either objective or applied in a blind fashion?

Yes.

• the composite outcome consisted of diagnoses that are objective made

If subgroups with different prognoses are identified, did adjustment for important prognostic factors take place?

Unclear, probably no.

• the authors provided results that were “adjusted for possible confounders” but did not explain which confounding factors
• it is highly probable that substantial residual confounding remains; the group receiving PPIs also were also at increased risk of PUD even with adjustment, contrary to data from randomised controlled trials
• this is another serious thread to internal validity

What were the results?

Primary outcome:

Composite outcome (myocardial infarction, unstable angina, stroke, all-cause mortality):

• PPI + clopidogrel group: 754 (13%)
• clopidogrel only group: 830 (7%)
• adjusted hazard ratio (HR), PPI + clopidogrel group vs clopidogrel-only group:
  • HR = 1.75 (95% CI 1.58-1.94)
  • interpretation: the PPI + clopidogrel group were at 75% higher risk of the primary outcome compared to the clopidogrel-only group
• mean time to the occurrence of cardiovascular event was 75 days

Other outcomes:

Complicated peptic ulcer disease, PPI + clopidogrel group vs clopidogrel-only group:

• HR = 4.76 (95% CI 1.18 – 19.17)

Will the results help me care for my patient?

External validity: Wikipedia

Are the participants different to my patient?

• older population; mean ages of 68.6 and 66.1 in the two groups studied
• roughly equal distribution between men and women
• approximately one third of patients had a history of myocardial infarction
• approximately one third of patients had a history of unstable angina pectoris
• the majority of patients were receiving: statins, β-blockers, ACE inhibitors, or aspirin, or a combination of these

Will this evidence make a clinically important impact on my conclusions about what to offer to tell my patients?

No.

• the evidence is too unreliable to be the basis of a change in management
• the risk, if any, of concurrent PPI in clopidogrel should be a reminder that PPIs should only be used if clinically indicated

Study weaknesses (summary)

• indication of what the authors term “channeling bias”; high likelihood that the concurrent PPI and clopidogrel group were less healthy
• the source database used did not contain information on important prognostic factors, e.g., smoking and obesity
  • the authors were unable to correct for these factors in their multivariate analysis
• it is probable that the magnitude of effect from the biases are substantially larger than studied effect (if it exists), invalidating the results

Study strengths (summary)

• first European cohort study investigating an association between concomitant use of clopidogrel and PPIs and the occurrence of cardiovascular events
• patients were censored after adverse events
• authors investigated the occurrence of GI events in this cohort, which was not done in the other similar studies
  • this gives an indication of the presence of unadjusted confounding
• large proportion (approximately 25%) of the Dutch population was included in the database

Biases and conflicts of interests

• van Boxel OS (primary author) was funded by an unrestricted grant from AstraZeneca (the manufacturer of omeprazole and esomeprazole)
• several other authors have served as consultants or received grants from various pharmaceutical companies including AstraZeneca
• the sponsors were not involved in the design, conduct, analysis, interpretation, review or approval of the study

Clinical relevance to primary health care

Clopidogrel (Plavix, Iscover) is widely used for the secondary prevention of cardiovascular disease in some settings (e.g., cerebrovascular disease, and myocardial infarction). To address the increased risk of gastrointestinal bleeding events in patients using clopidogrel, the American College of Cardiology recommends the use of prophylactic proton pump inhibitors (PPIs) in at-risk patients ^[2].

There is basic pharmacokinetic data that suggests that some PPIs decrease the actions of clopidogrel and several observational studies, including the described study, have demonstrated significant associations of harm.  However, in a recent trial, meta-analysis of randomised patients found no association of excess cardiovascular risk or overall mortality with concurrent PPI use ^[3]. The meta-analysis was limited by significant heterogeneity.

Australian general practitioners should consider the evidence base of increased cardiovascular risk from the use of PPIs in patients taking clopidogrel to be unclear. The nature of observational studies in general limits the causative statements that can be made. The results of this study in particular are probably invalid; there are indications that large confounding effects remain (even after adjustments, use of PPIs was still associated with almost a 5-fold increased risk of complicated peptic ulcer disease, completely contrary to the known empiric effects of PPIs and basic pathophysiology).

In the absence of definitive evidence, primary care physicians should be guided by a precautionary principle. It is reasonable to treat patients at high risk of gastrointestinal bleeding, or have other clear indications, with PPIs. The need for starting or ongoing PPIs should be assessed from the perspective that PPIs might increase the risk of cardiovascular outcomes ^[4].

References

1. van Boxel OS, van Oijen MGH, Hagenaars MP, et al. Cardiovascular and gastrointestinal outcomes in clopidogrel users on proton pump inhibitors: results of a large Dutch cohort study.  Am J Gastroenterol 2010; 105: 2430–6
2. Bhatt DL , Scheiman J, Abraham NS, et al. 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 Clinical Expert Consensus Documents. Circulation 2008; 118: 1894–909.
3. Kwok CS, Loke YK. Meta-analysis: the effects of proton pump inhibitors on cardiovascular events and mortality in patients receiving clopidogrel. Aliment Pharmacol Ther 2010; 31(8): 810–23
4. NPS Prescribing practice review 45: proton pump inhibitors: step-down to symptom control. National Prescribing Service Limited. 4 May 2009

Co-author: Michael Tam