Use of Acid-Suppressing Drugs Before the Occurrence of Ischemic Events
Use of Acid-Suppressing Drugs Before the Occurrence of Ischemic Events
Study Objective. To determine if the use of acid-suppressing drugs is increased before the occurrence of ischemic events.
Design. Population-based, nested case-control analysis.
Data Source. Administrative databases in Saskatchewan, Canada.
Patients. Cases were 1612 patients (aged ≥ 40 yrs) who started a first-ever antihypertensive drug between January 1, 1994, and December 31, 2003, and were hospitalized for a first ischemic heart event of either myocardial infarction (1002 patients) or unstable angina (610 patients); five control patients were matched to each case patient by age, sex, and year of first antihypertensive prescription (8060 controls).
Measurements and Main Results. Within the case and control groups, we calculated exposure to acid-suppressing therapy, defined as proton pump inhibitors (PPIs) or histamine2-receptor antagonists (H2RAs), within 90 days leading up to the event. Exposure to acid-suppressing therapy was higher among cases than controls (15.3% [246/1612] vs 10.4% [837/8060], adjusted odds ratio [AOR] 1.26, 95% confidence interval [CI] 1.06–1.49, p<0.009). Exposure to each acid suppressant was similarly higher among cases than controls: H2RA users (11.7% [188/1612] vs 8.4% [678/8060], AOR 1.21, 95% CI 1.00–1.46, p<0.048) and PPI users (4.0% [64/1612] vs 2.2% [179/8060], AOR 1.32, 95% CI 0.95–1.84, p=0.094). Use of other drugs was also significantly increased during this period.
Conclusions. Use of acid-suppressing drugs increased before the occurrence of ischemic events regardless of the type (PPI or H2RA) or whether other drugs, such as clopidogrel, were concurrently administered. In addition, significant increases in overall drug use were observed during this time frame, suggesting that many patients exhibit warning signs before an acute hospitalization. Thus, PPI use before the occurrence of ischemic events may simply be a marker of unmeasured and uncontrolled confounding in observational studies that have implicated a PPI-clopidogrel interaction as a cause of recurrent ischemic events.
In 2009, two observational studies raised serious concern about the potential for adverse outcomes in patients taking proton pump inhibitors (PPIs) concurrently with clopidogrel. This apparent drug interaction was attributed to an inhibitory effect of PPIs on cytochrome P450 2C19 isoenzymes, which subsequently prevents the bioactivation of clopidogrel. Laboratory studies have demonstrated that certain PPIs such as lansoprazole and omeprazole inhibit the antiplatelet effects of clopidogrel, whereas pantoprazole and esomeprazole do not.
Subsequent studies have attempted to verify the original observations of adverse events associated with combining PPIs and clopidogrel. However, a clear answer remains elusive because the findings have been inconsistent. Although the studies are conflicting overall, it appears that the null hypothesis—that no harm results from the drug combination—is generally favored among the highest quality studies. Indeed, randomized controlled trials, post hoc analyses of randomized controlled trials, or propensity matched studies have not shown a significant clinical impact of using clopidogrel with PPIs.
Although a specific source of confounding has not been identified, we believe that observational studies in this setting are inherently confounded because PPI use likely increases before the occurrence of ischemic events regardless of the use of clopidogrel. The link between acid-suppressing drugs and ischemic events is likely mediated by perceived symptoms of dyspepsia or gastroesophageal reflux disease (GERD) that are actually a result of anginal symptoms. Thus, it is more likely that PPI use before ischemic events is simply a marker of unmeasured and uncontrolled confounding in the observational data. Indeed, one group of investigators reported that PPIs were not associated with ischemic events on their own; however, examination of the survival data revealed two clearly distinct patterns that challenge their conclusions. First, PPI users (compared with nonusers) were actually at a greater risk during the first year, and second, PPI use alone was associated with a higher risk of ischemic events compared with PPIs combined with clopidogrel. As there is little clinical or biologic evidence to suggest PPIs alone contribute to ischemic events, it is more likely that PPI use represented unmeasured and uncontrolled confounding. Unfortunately, the use of histamine2-receptor antagonists (H2RAs) as monotherapy was not examined in that study; use of these drugs could have helped disentangle the role of confounding (or protopathic bias) from true adverse drug events.
Another group of investigators also reported an increase in recurrent myocardial infarction among elderly patients taking PPIs and clopidogrel concurrently in Ontario, Canada. Unlike the previous study, PPI use alone was not associated with a higher risk for events. In addition, the investigators found no association between H2RAs and myocardial infarction; however, it was not clear if H2RA use was evaluated in the same manner as PPIs. For instance, previous PPI exposure was grouped into three categories (0–30, 31–90, and 91–180 days) in relation to readmission for myocardial infarction, whereas exposure to H2RAs was only examined as a single category. Also, H2RA use was very low in both groups, resulting in a very large confidence interval (CI 0.63–1.40) surrounding the odds ratio (OR) for recurrent events.
Presumably, if confounding (or protopathic bias) substantially contributes to the previously observed association between PPI-clopidogrel combinations and recurrent myocardial infarction in observational data, a similar pattern of association should exist with H2RAs that have inherently no biologic mechanism for an interaction. We hypothesized that the use of all acid-suppressing therapies increases before the occurrence of ischemic events irrespective of the use of clopidogrel. Specifically, we examined a population-based cohort of newly treated patients with hypertension to evaluate the use of any acid-suppressing therapy during the period preceding a first ischemic heart event.
Abstract and Introduction
Abstract
Study Objective. To determine if the use of acid-suppressing drugs is increased before the occurrence of ischemic events.
Design. Population-based, nested case-control analysis.
Data Source. Administrative databases in Saskatchewan, Canada.
Patients. Cases were 1612 patients (aged ≥ 40 yrs) who started a first-ever antihypertensive drug between January 1, 1994, and December 31, 2003, and were hospitalized for a first ischemic heart event of either myocardial infarction (1002 patients) or unstable angina (610 patients); five control patients were matched to each case patient by age, sex, and year of first antihypertensive prescription (8060 controls).
Measurements and Main Results. Within the case and control groups, we calculated exposure to acid-suppressing therapy, defined as proton pump inhibitors (PPIs) or histamine2-receptor antagonists (H2RAs), within 90 days leading up to the event. Exposure to acid-suppressing therapy was higher among cases than controls (15.3% [246/1612] vs 10.4% [837/8060], adjusted odds ratio [AOR] 1.26, 95% confidence interval [CI] 1.06–1.49, p<0.009). Exposure to each acid suppressant was similarly higher among cases than controls: H2RA users (11.7% [188/1612] vs 8.4% [678/8060], AOR 1.21, 95% CI 1.00–1.46, p<0.048) and PPI users (4.0% [64/1612] vs 2.2% [179/8060], AOR 1.32, 95% CI 0.95–1.84, p=0.094). Use of other drugs was also significantly increased during this period.
Conclusions. Use of acid-suppressing drugs increased before the occurrence of ischemic events regardless of the type (PPI or H2RA) or whether other drugs, such as clopidogrel, were concurrently administered. In addition, significant increases in overall drug use were observed during this time frame, suggesting that many patients exhibit warning signs before an acute hospitalization. Thus, PPI use before the occurrence of ischemic events may simply be a marker of unmeasured and uncontrolled confounding in observational studies that have implicated a PPI-clopidogrel interaction as a cause of recurrent ischemic events.
Introduction
In 2009, two observational studies raised serious concern about the potential for adverse outcomes in patients taking proton pump inhibitors (PPIs) concurrently with clopidogrel. This apparent drug interaction was attributed to an inhibitory effect of PPIs on cytochrome P450 2C19 isoenzymes, which subsequently prevents the bioactivation of clopidogrel. Laboratory studies have demonstrated that certain PPIs such as lansoprazole and omeprazole inhibit the antiplatelet effects of clopidogrel, whereas pantoprazole and esomeprazole do not.
Subsequent studies have attempted to verify the original observations of adverse events associated with combining PPIs and clopidogrel. However, a clear answer remains elusive because the findings have been inconsistent. Although the studies are conflicting overall, it appears that the null hypothesis—that no harm results from the drug combination—is generally favored among the highest quality studies. Indeed, randomized controlled trials, post hoc analyses of randomized controlled trials, or propensity matched studies have not shown a significant clinical impact of using clopidogrel with PPIs.
Although a specific source of confounding has not been identified, we believe that observational studies in this setting are inherently confounded because PPI use likely increases before the occurrence of ischemic events regardless of the use of clopidogrel. The link between acid-suppressing drugs and ischemic events is likely mediated by perceived symptoms of dyspepsia or gastroesophageal reflux disease (GERD) that are actually a result of anginal symptoms. Thus, it is more likely that PPI use before ischemic events is simply a marker of unmeasured and uncontrolled confounding in the observational data. Indeed, one group of investigators reported that PPIs were not associated with ischemic events on their own; however, examination of the survival data revealed two clearly distinct patterns that challenge their conclusions. First, PPI users (compared with nonusers) were actually at a greater risk during the first year, and second, PPI use alone was associated with a higher risk of ischemic events compared with PPIs combined with clopidogrel. As there is little clinical or biologic evidence to suggest PPIs alone contribute to ischemic events, it is more likely that PPI use represented unmeasured and uncontrolled confounding. Unfortunately, the use of histamine2-receptor antagonists (H2RAs) as monotherapy was not examined in that study; use of these drugs could have helped disentangle the role of confounding (or protopathic bias) from true adverse drug events.
Another group of investigators also reported an increase in recurrent myocardial infarction among elderly patients taking PPIs and clopidogrel concurrently in Ontario, Canada. Unlike the previous study, PPI use alone was not associated with a higher risk for events. In addition, the investigators found no association between H2RAs and myocardial infarction; however, it was not clear if H2RA use was evaluated in the same manner as PPIs. For instance, previous PPI exposure was grouped into three categories (0–30, 31–90, and 91–180 days) in relation to readmission for myocardial infarction, whereas exposure to H2RAs was only examined as a single category. Also, H2RA use was very low in both groups, resulting in a very large confidence interval (CI 0.63–1.40) surrounding the odds ratio (OR) for recurrent events.
Presumably, if confounding (or protopathic bias) substantially contributes to the previously observed association between PPI-clopidogrel combinations and recurrent myocardial infarction in observational data, a similar pattern of association should exist with H2RAs that have inherently no biologic mechanism for an interaction. We hypothesized that the use of all acid-suppressing therapies increases before the occurrence of ischemic events irrespective of the use of clopidogrel. Specifically, we examined a population-based cohort of newly treated patients with hypertension to evaluate the use of any acid-suppressing therapy during the period preceding a first ischemic heart event.
Source...