Prospective, Randomized, Placebo-Controlled Evaluation of Esomeprazole in Coronary Artery Disease Patients. EPAC: Esomeprazole Prevention of Atypical Chest Pains

ABSTRACT: Background. Coronary artery disease (CAD) patients experience chest pain (CP) from cardiac and non-cardiac etiologies. Objective. Proton pump inhibitor (PPI) treatment of CAD patients with atypical CP would result in less CP, as well as fewer emergency room (ER) visits for CP, medical evaluations, and hospitalizations. Methods. A randomized, prospective, placebo-controlled trial in patients with a history of severe CAD, without current ischemia, and not receiving acid reducing therapy. Patients received esomeprazole 40 mg or placebo daily added to standard cardiac medications for 6 months. The primary endpoint was percentage of patients experiencing CP. Results. In 162 patients randomized, 78 received esomeprazole and 84 received placebo. Esomeprazole significantly reduced the primary endpoint of patients experiencing CP (24.4% versus 54.8%; p < 0.001). The absolute number of CP episodes and CP days was also significantly reduced. Fewer patients experienced the combined endpoint of doctor office/ER visits for CP (19.2% versus 48.8%; p < 0.001), mainly due to significantly fewer office visits for CP (14.1% versus 39.3%; p < 0.001). ER visits for CP trended less with esomeprazole (12.8% versus 22.6%; p = 0.10), as did hospitalizations for CP (14.1% versus 20.2%; p = 0.30). However, the study lacked statistical power to reach these secondary endpoints. Conclusions. Concomitant esomeprazole therapy in CAD patients without active ischemia but with a history of atypical CP reduces number of patients with further CP symptoms and medical office visits for CP. Larger studies are needed to further evaluate PPI treatment on ER CP visits and hospitalizations in this patient population. 

J INVASIVE CARDIOL 2011;23:222–226

Key words: chest pain, coronary disease, emergency department, hospitalization, proton pumps

__________________________________________

Coronary artery disease (CAD) patients frequently suffer from chest discomfort both of coronary etiology secondary to ischemia and non-cardiac CP. It is often clinically difficult during outpatient and inpatient evaluations to delineate between cardiac and gastrointestinal causes.^1,2 Considering all patients who visit the emergency room (ER) with a primary complaint of CP, over 80% may not get a final diagnosis of cardiac ischemia.^3,4 In high-risk unstable angina patients, 10–40% do not have ischemic CAD.^5–7 Among CAD patients experiencing CP, 50% have gastroesophageal reflux disease (GERD) as a cause of non-cardiac CP.⁸ Retrospective analysis showed proton pump inhibitor (PPI) associated with fewer patients experiencing CP, ER visits, and hospitalizations, as well as fewer of these adverse events.⁹

We sought to prospectively evaluate whether treatment with PPI therapy for previously symptomatic CAD patients, without known gastrointestinal disease and without proven coronary ischemia, would affect the incidence of future CP episodes. PPI therapy will be evaluated when added to concomitant cardiac medical therapy. No prospective studies to date have evaluated such treatments to prevent CP in patients with CAD. It was hypothesized that CAD patients utilizing PPI therapy would have reduced CP episodes, as well as fewer ER presentations and hospital admissions for CP.

Methods

Patient population. The study was performed at a single, tertiary-care medical center. Patients were enrolled within a period of 1 year. All patients were informed participants and signed consent forms approved by the Institutional Review Board.

Inclusion criteria for a history of CAD required one of the following: documented myocardial infarction (MI) by Q-waves present in two or more contiguous leads on electrocardiogram (ECG) or elevation of cardiac enzymes in accordance with World Health Organization guidelines;¹⁰ a coronary angiogram with obstructive CAD defined as > 50% luminal narrowing; a revascularization history by percutaneous coronary intervention or coronary artery bypass graft surgery; or history of an abnormal stress test demonstrating infarction or ischemia (the latter already corrected with revascularization). All enrolled patients must have atypical CP symptoms experienced within the past year. Atypical CP symptoms were defined as non-exertional, or CP at rest, not exacerbated by stress, not relieved with nitroglycerin or rest, of brief duration lasting seconds to a few minutes, or of prolonged duration lasting many hours, as well as CP that was not substernal, e.g., right-sided. Symptoms needed to be deemed non-ischemic by the patient’s physician. Patients had an appropriate cardiac stress testing and/or cardiac catheterization with or without revascularization prior to enrollment.

Exclusion criteria were any one of the following: 1) patients without a history of CAD; 2) CAD patients not experiencing symptoms; 3) patients taking PPI or H2-antagonist therapy within 1 month; 4) patients with current non-invasive cardiac evaluations that revealed coronary ischemia yet to be revascularized; 5) patients with pre-existing severe liver disease; 6) patients taking medications of which PPI interfered with the absorption, including iron tablets, anti-retroviral agents, anti-fungals, and cephalosporins. Patients with indications for long-term use of stomach acid reducing agents, e.g., gastrointestinal bleed, severe gastritis, peptic ulcer disease, or GERD diagnosed by a gastroenterologist, were not candidates for enrollment.

Patient enrollment. Computer-generated randomization was performed in a 1:1 ratio, esomeprazole to placebo, in blocks of 50. Two non-investigator volunteers, blinded to subsequent study enrollment, accessed the randomization scheme and labeled the respective study drug bottles. Study drug was packaged with corresponding numbered monthly diary cards and 6 return-addressed envelopes with return postage, and distributed to enrolled subjects. Patients were given 6 months of medication divided among 6 bottles, 35 capsules per bottle, of either esomeprazole 40 mg or identical-appearing placebo capsules. Subjects, caregivers, and investigators were blinded to treatment strategy throughout the study. Esomeprazole and placebo were donated by Astra Zeneca LP (Sweden).

Measurement of outcomes. Baseline demographic information, such as age, gender, race, and history of hypertension, hypercholesterolemia, diabetes mellitus, MI, PCI, CABG, coronary heart failure, stroke, peripheral vascular disease, current tobacco use (within the prior 1 year), alcohol use (> 2 drinks/week), exercise habits (active defined as > 3x per week), and use of common cardiovascular medications taken concomitantly during the study were recorded.

Recorded outcomes included: 1) incidence of CP events; 2) outpatient visits to a medical doctor (MD) for CP; 3) ER visits for CP; and 4) hospitalization for a primary complaint of CP. For patients who sought medical help for CP (MD visits, ER visits, or hospitalization), additional telephone follow-ups were performed. Investigators confirmed that events were primarily for CP, and elicited whether the events were primarily cardiac or non-cardiac in etiology. ECGs, cardiac enzymes, echocardiograms, stress tests, cardiac catheterizations, and PCI information were recorded when applicable.

Endpoints. The primary endpoint was the percentage of patients experiencing CP during 6 months of follow-up. Secondary endpoints assessed were percentage of patients with: MD office/ER visits for CP (combined endpoint), MD office visits for CP, ER visits for CP, hospitalizations for CP, and the cumulative number of CP episodes, symptom-free days, MD office/ER visits for CP and hospitalizations for CP. A cardiac etiology of symptoms was met by the following criteria: evidence of new abnormal ECG findings, e.g., ST depressions at rest or during exercise tests; evidence of new cardiac enzyme release (2 serial samples); evidence of reversible perfusion abnormality on nuclear stress; evidence of new obstructive coronary lesion by angiography; or occurrence of PCI on a culprit artery. A non-cardiac etiology of symptoms was met by the following criteria: no new ECG abnormalities; no abnormal cardiac enzymes; a nuclear stress test without ischemia; or coronary angiogram without new abnormalities, as well as the patient’s cardiologist’s declaration that cardiac condition was stable, thereby informing them to continue medical therapy because symptoms were not likely to be cardiac in origin.

Number needed to treat. This was a pilot study. There were no published data demonstrating number of episodes of CP that patients with CAD have per year, nor data suggesting how many ER or MD office visits for CP this population may have per year. The prior incidence of events for patients with CAD at our institution was also unknown. We calculated the number needed for this study using published estimates that 7.5 million people per year present to the ER with CP. Assuming most CAD patients are above age 55, and using United States census estimates for year 2000 of 63 million persons above age 55, this translated into an incidence proportion of 12%. A population with established CAD likely would have a higher percentage of people presenting to the ER for CP (both cardiac and non-cardiac CP). Nevertheless, using the above estimate, a 50% reduction in percentage of people presenting to the ER for CP (any type) with 80% power and a two-sided alpha level of 5%, would require 390 patients per arm (780 total). An 80% reduction would require 133 patients per arm (266 total).

Statistical and data analyses. The measured variables and outcomes for each patient were entered into SPSS for Windows, version 10.0.1 (SPSS Inc., Chicago, Illinois) for statistical analyses. Data were analyzed by the intention-to-treat principle. Continuous variables were expressed as mean ± standard error of the mean (SE) and compared by unpaired student t-test. Nominal or dichotomous variables are expressed as proportions, and compared by Chi-squared test. Ordinal values, such as the number of CP events, ER visits and hospitalizations, were analyzed using non-parametric testing based on rank sums using the Mann-Whitney test. p-values < 0.05 were considered statistically significant.

Results

All patients enrolled had CAD and no patients were taking PPI or H2-antagonists. Of the 162 patients enrolled, 95% of patients completed the study at 6 months of follow-up (5 placebo, 3 esomeprazole patients were lost to follow-up prior to intention-to-treat endpoint). Twenty-five patients (9 placebo; 16 esomeprazole; p = 0.09) terminated treatment earlier than 6 months due to adverse drug effects in 8 patients (1 placebo, 7 esomeprazole), primary physician preference in 6 patients, non-compliance in 7 patients, gastrointestinal bleed in 1 patient, gastritis in 1 patient, and unknown reason in 2 patients. Nineteen patients (7 placebo, 12 esomeprazole; p = 0.16) experienced mild adverse effects, including: rash (3.1%); diarrhea (1.9%); nausea/vomiting (1.9%), constipation (1.2%); abdominal pain (1.2%); body aches (0.6%); flatus (0.6%); headache (0.6%); and facial swelling (0.6%). The incidences of these adverse effects are consistent with those recorded in the package insert of the study drug.

Baseline features of the 162 enrolled CAD patients separated by their randomized groupings are shown in Table 1. There were no differences between esomeprazole and placebo groups with regard to patient age, race, gender, or history of hypertension, diabetes, hypercholesterolemia, prior MI, prior PCI, CABG, CHF, cerebrovascular accident, PVD, smoking, alcohol use, exercise habits, or evidence of an obese phenotype. There were no significant differences in relevant medications used between esomeprazole and placebo-treated patients (Table 2). There was a high utilization of standard CAD medications, e.g.,  aspirin, clopidogrel, beta-blocker, and/or statin, implying that enrolled patients were closely adherent to recommended treatment guidelines.

CAD patients who received daily esomeprazole 40 mg versus placebo for 6 months had a significantly lower incidence of the primary endpoint of CP (24.3% versus 54.8%; p < 0.001) (Table 3). This represented a 55% relative reduction in the number of patients with CP if on esomeprazole. Addition of esomeprazole translated into a 79% relative increase in symptom-free patients. Interestingly, among patients who experienced CP, 71% were not taking PPI therapy.

The combined endpoint of visits to an MD’s office or ER for the primary complaint of CP was reduced by esomeprazole (48.8% versus 19.2%; p < 0.001). This translated into a 61% reduction (Table 3). The result was driven primarily by significantly fewer MD office visits (39.3% versus 14.1%; p <0.001; 64% relative reduction). A trend toward fewer ER visitations was observed with esomeprazole versus placebo (22.6% versus 12.8%; p = 0.10). MD/ER visits after placebo treatment were 85% non-cardiac (35/41 patients) and 15% cardiac (6/41 patients), versus 27% non-cardiac (4/15 patients) and 73% cardiac (11/15 patients) with esomeprazole (p < 0.001).

Patients presenting to the ER or MD office for CP were susceptible to hospitalization for further evaluation of their symptoms (Table 3). 20.2% of placebo-treated patients versus 14.1% of esomeprazole-treated patients were hospitalized for CP (p = 0.30). Hospitalizations after placebo treatment were 82% non-cardiac (14/17 patients) and 18% cardiac (3/17 patients), versus 27% non-cardiac (3/11 patients) and 73% cardiac (8/11 patients) with esomeprazole (p = 0.007).

The number of documented CP events was also higher in those receiving placebo versus esomeprazole. There were 515 CP episodes among 79 placebo-treated patients, compared to 229 CP episodes among the 75 esomeprazole-treated patients, representing a 53% reduction (p < 0.001) with esomeprazole (Table 4). The cumulative number of days patients experienced CP was elevated in placebo- versus esomeprazole-treated patients (463 versus 224; p < 0.001), and the number of CP days/patient over 6 months was 5.9 versus 3.0 days for placebo and esomeprazole patients, respectively. Number of ER and MD visits for CP was reduced by 68% (76 versus 23) in patients taking esomeprazole (p < 0.001). This translated into 0.96 and 0.31 MD/ER visits/patient over 6 months for placebo versus esomeprazole treatment.

Discussion

This paper is the first demonstration in a randomized, prospective, placebo-controlled trial that PPI therapy, when added to concomitant cardiac medical therapy in CAD patients, effectively reduces symptoms of CP. It is believed that PPI therapy reduced CP caused by GERD. Esomeprazole is an effective and approved treatment for GERD. However, the treatment of atypical or non-cardiac CP is not a current Food and Drug Administration-approved indication. In this study, esomeprazole was able to reduce the number of patients experiencing CP, and increase the number of symptom-free days. Esomeprazole reduced the combined endpoint of patient ER/physician office visits for CP. Although ER visits alone and hospitalizations for CP trended lower with esomeprazole, results were not statistically significant, likely secondary to the small study population. We believe the reduction in CP remains clinically important from a quality-of-life standpoint.

A history of significant CAD leads to hospital evaluations for CP, since this history alone risk stratifies a patient to a diagnosis of unstable angina with a high likelihood that their symptoms represent an acute coronary syndrome.^11,12 In clinical trial data, some patients diagnosed with unstable angina have normal coronary arteries or insignificant disease, making it possible that some symptoms were non-cardiac in this subgroup.^5–7,13,14 GERD is a common cause of non-cardiac CP in patients both with and without CAD.⁸ GERD is prevalent; in United States epidemiological studies, reflux symptoms exist in 1 in 5 adults (20%) with weekly symptoms, and 10–40% have monthly symptoms.¹⁵ It is reasonable that esomeprazole significantly lowers adverse events of CP symptoms in CAD patients, especially since atypical angina pectoris and GERD-induced non-cardiac CP can be clinically indistinguishable.¹⁶

In this study, assessment of CP etiologies causing MD/ER visits and hospitalizations by appropriate CP work-ups revealed etiologies to be predominantly cardiac for esomeprazole patients consistent with a treatment effect for GERD. Approximately 25% of CP etiologies evaluated in PPI treated patients were non-cardiac, whereas approximately 85% of symptoms in placebo-treated patients were non-cardiac. Etiologies were primarily non-cardiac for placebo-treated patients demonstrating lack of PPI treatment effect. This is the first demonstration of a reduction in symptoms and evaluations, specifically non-cardiac CP evaluations, in a CAD population with a drug utilized for GERD.

The results of this study show potential benefit for CAD patients with a history of CP symptoms. It is impossible at this time to predict which patients will experience non-cardiac CP from GERD, as it is impossible to predict who may experience an acute myocardial infarction. Multiple pharmacologic agents prove beneficial to reduce cardiac angina pectoris, e.g., aspirin, beta-blockers, nitrates, calcium-channel blockers, and statins.¹⁷ These data suggest esomeprazole therapy might be utilized in addition to, not as a replacement to, standard cardiac medications.

The limitations of this study include the fact that it is a single-center study, creating possible institutional bias. The relatively small sample size, although large for a single center, did not reach statistical significance for the endpoints of ER visits alone or overall hospitalizations for CP. The EPAC study was powered to a secondary endpoint of ER visits, which was not met, although the primary endpoint in this hypothesis-generating study was statistically significant. A larger, multicenter study in selected patients with CAD evaluating the endpoints of ER visits or hospitalizations would be required to clarify the issue of whether PPI therapy, or specifically esomeprazole, can reduce these events.

In summary, CAD patients with atypical CP appear to have effective relief from daily PPI therapy with minimal adverse effects. Prescribed esomeprazole therapy significantly associates with a reduced number of important outcomes, specifically, reduced future onset of CP symptoms and reduced evaluations for CP, as well as increased number of symptom-free days.

References

1. Schappert SM. National Hospital Ambulatory Medical Care Survey: 1992 emergency department summary. Vital Health Stat 13 1997;125:1–108.
2. Hutter AM, Amsterdam EA, Jaffe AS. Acute coronary syndromes: Chest discomfort evaluation in the hospital. J Am Coll Cardiol 2000;35:853–862.
3. Udelson JE, Beshansky JR, Ballin DS, et al. Myocardial perfusion imaging for evaluation and triage of patients with suspected acute cardiac ischemia. A randomized controlled trial. JAMA 2002;288:2693–2700.
4. Katz PO, Castell DO. Approach to the patient with unexplained chest pain. Am J Gastroenterol 2000;95:S4–S8.
5. Diop D, Aghababian RV. Definition, classification, and pathophysiology of acute coronary ischemic syndromes. Emerg Med Clin North Amer 2001;19:259–267.
6. Pope JH, Ruthazer R, Beshansky JR, et al. Clinical features of emergency room department patients presenting with symptoms suggestive of acute cardiac ischemia. J Thromb Thrombolysis 1998;6:63–74.
7. Kontos MC, Ornato JP, Schmidt JL, et al. Incidence of high-risk acute coronary syndromes and eligibility for glycoprotein IIb/IIIa inhibitors among patients admitted for possible myocardial ischemia. Am Heart J 2002;143:70–75.
8. Liuzzo JP, Ambrose JA. Chest pain from gastroesophageal reflux disease in patients with coronary artery disease. Cardiol Rev 2005;13:167–173.
9. Liuzzo JP, Ambrose JA, Diggs P. Proton pump inhibitors for patients with coronary artery disease associated with reduced chest pain, emergency department visits, and hospitalizations. Clin Cardiol 2005;28:369–374.
10. Luepker RV, Apple FS, Christenson RH, et al. AHA Scientific Statement. Case definitions for acute coronary heart disease in epidemiology and clinical research studies. Circulation 2003;108:2543–2549.
11. Braunwald E. Unstable angina: A classification. Circulation 1989;80:410–414.
12. Braunwald E, Antman EM, Beasley JW. ACC/AHA guidelines for the management of patients with unstable angina/non-ST segment elevation myocardial elevation. J Am Coll Cardiol 2000;36:970–1062.
13. Van Miltenberg AJ, Simoons ML, Veerhoek RJ, Bossuyt PM. Incidence and follow-up of Braunwald subgroups of unstable angina. J Am Coll Cardiol 1995;25:1286–1292.
14. Isreal DH, Ambrose JA. Angiographic correlation of complex lesions in acute coronary syndromes. In: Complex Lesions in Acute Coronary Syndromes. Ambrose J (ed). New York, New York: Future Publishing.1996: pp. 49–72.
15. Locke GR, Talley NJ, Fett SL, et al. Prevalence and clinical spectrum of gastroesophageal reflux: A population-based study in Olmsted County, Minnesota. Gastroenterology 1997;112:1148–1156.
16. Singh S, Richter J, Hewson E, et al. The contribution of gastroesophageal reflux to chest pain in patients with coronary artery disease. Ann Intern Med 1992;117:824–830.
17. Gibbons R, Abrams J, Chatterjee K, et al. ACC/AHA 2002 guideline update for the management of patients with chronic stable angina. Available at www.acc.org/clinical/guidelines/stable/stable.pdf

_________________________________________

From Saint Vincent Catholic Medical Centers, New York, New York.
This work was supported by an investigator-sponsored trial research grant from AstraZeneca LP to the institution.
Manuscript submitted February 28, 2011, provisional acceptance given March 15, 2011, final version accepted April 4, 2011.
Address for correspondence: John P. Liuzzo, MD, PhD, Hackensack University Medical Center, Mulkay Cardiology Consultants, PC, 529 39 St., Union City, NJ 07087. Email: jliu2006@yahoo.com