Author + information
- Received February 18, 1998
- Revision received July 23, 1998
- Accepted August 6, 1998
- Published online December 1, 1998.
- Karen P Alexander, MD∗,* (, )
- Eric D Peterson, MD, MPH∗,
- Christopher B Granger, MD, FACC∗,
- A.Cecilia Casas, MS∗,
- Frans Van de Werf, MD†,
- Paul W Armstrong, MD, FACC‡,
- Alan Guerci, MD, FACC§,
- Eric J Topol, MD, FACC∥,
- Robert M Califf, MD, FACC∗,
- for the GUSTO-IIB Investigators
- ↵*Address for correspondence: Karen P. Alexander, MD, Duke Clinical Research Institute, 2400 Pratt Street, Suite 7068, Durham, North Carolina 27705
Objectives. The purpose of this study to determine whether use of cardiac medications reflects evidence-based recommendations for patients with non–ST elevation acute coronary syndromes.
Background. Agency for Health Care Policy and Research practice guidelines for unstable angina recommend the use of cardiac medications based on evidence from randomized trials. It is unknown whether practitioners in the U.S., Canada and Europe follow these recommendations in patients with non–ST elevation acute coronary syndromes.
Methods. We studied 7,743 patients with non–ST elevation acute coronary syndromes enrolled in the international Global Use of Strategies to Open Occluded Arteries in Acute Coronary Syndromes trial. The use of aspirin, beta-adrenergic blocking agents, angiotensin-converting enzyme inhibitors and calcium channel blocking agents was determined at discharge for all patients and “ideal” patients (those with indications and no contraindications). Using published estimates of relative mortality reductions with these drugs, we calculated the lives that could have been saved at 1 year if discharge medication use had better matched guideline recommendations.
Results. Overall, guideline adherence at discharge in “ideal” patients was 85.6% for aspirin, 59.1% for beta-blockers and 51.7% for angiotensin-converting enzyme inhibitors. Calcium channel blockers were given to 26.7% of patients with a contraindication to these drugs. These rates were similar across locations of enrollment. Women and older patients less often received aspirin when “ideal,” and younger patients more often received calcium channel blockers when they were contraindicated. If medication use had been more evidence-based, 1-year mortality might have been reduced by a relative 22%.
Conclusions. There is significant room for improvement in the use of recommended drugs in patients with non–ST elevation acute coronary syndromes. Medication use that more closely follows recommendations could reduce mortality in this population.
“Unstable angina” describes a broad spectrum of non–ST elevation acute coronary syndromes, including crescendo angina, angina at rest and non–Q wave myocardial infarction. Across this range of clinical syndromes, unstable angina is a frequent cause of morbidity and has been estimated to have a 1-year mortality as high as 12%, most of this risk occurring within the first months after hospital discharge (1–3). Randomized trials have provided clear evidence that certain medications can reduce recurrent ischemic events and improve survival in patients with non–ST elevation acute coronary syndromes. The Agency for Health Care Policy and Research (AHCPR) and the National Heart, Lung, and Blood Institute Clinical Practice Guidelines for the Treatment of Unstable Anginasummarize this knowledge and recommend treatments shown to be beneficial in clinical trials, in an attempt to influence practice patterns and improve outcomes (1). Prescribing patterns that fall short of guideline recommendations may adversely affect the survival of patients with acute coronary syndromes.
Our analysis describes the use of guideline-recommended discharge medications in patients with non–ST elevation acute coronary syndromes enrolled in the international Global Use of Strategies to Open Occluded Arteries in Acute Coronary Syndromes (GUSTO-IIb) trial. We investigated the use of four drugs (aspirin, beta-adrenergic blocking agents, angiotensin-converting enzyme inhibitors and calcium channel blocking agents) overall and in patients with indications but no contraindications to each therapy. We examined their use by place of enrollment and by predictors of appropriate care. Using estimates from the literature for baseline mortality in this population, we calculated the reduction in mortality that could have been achieved with a more evidence-based approach to the use of these drugs in patients presenting with unstable angina. Although medical regimens must continue to be tailored to the individual circumstances of the patient, our study identifies populations who may benefit from each therapy, enabling us to quantify the potential impact of closer guideline adherence.
The GUSTO-IIb trial has been described (4). Briefly, 12,142 patients with acute coronary syndromes were enrolled at 373 hospitals in 13 countries between May 1994 and October 1995. Patients were eligible for enrollment if they had chest pain within the previous 12 h associated with either transient or persistent electrocardiographic changes. The trial population included 4,131 patients with ST–segment elevation myocardial infarction and 8,011 with non–ST elevation acute coronary syndromes (unstable angina or non–Q wave myocardial infarction). Patients were randomized to receive 72 h of intravenous heparin or hirudin therapy in addition to standard treatment and were followed for a primary end point of death or myocardial (re)infarction at 30 days. Our analysis included the 7,743 patients with non–ST elevation acute coronary syndromes enrolled in GUSTO-IIb who survived to hospital discharge: 2,183 from the U.S., 892 from Canada and 4,668 from Europe.
This observational analysis investigated the use of aspirin, beta-blockers, angiotensin-converting enzyme inhibitors and calcium channel blockers in selected patients at discharge. Inclusion and exclusion criteria for each medication were derived from the Clinical Practice Guidelines for Unstable Angina, the American Heart Association/American College of Cardiology Practice Guidelines for Acute Myocardial Infarction and from established, validated quality indicators of medication use (1,5,6). These criteria were selected to exclude patients from the “ideal” category when clinicians might have disagreed about the appropriateness of the therapy, and may still include patients for whom therapy remains controversial. We screened the baseline characteristics and in-hospital complications available on the case report form for these criteria. The number of patients in our population meeting at least one inclusion or exclusion criterion for each medication, as well as the percentage of those patients qualifying by each criterion, are shown in the Appendix. The construction of the “ideal” patient, different for each medication, is also shown in the Appendix. First, potential patients were defined as those with inclusion criteria for the use of aspirin, beta-blockers, angiotensin-converting enzyme inhibitors or calcium channel blockers. “Ideal” patients were then obtained by removing patients with exclusion criteria from the population of potential patients, leaving only those who met inclusion criteria for use without meeting any exclusion criterion. For example, patients were considered “ideal” for angiotensin-converting enzyme inhibitor use if they had an ejection fraction <40% or in-hospital congestive heart failure but did not have chronic renal insufficiency, a creatinine level >177 μmol/L (>2.0 mg/dL) or an ejection fraction >40%. For aspirin, beta-blockers and angiotensin-converting enzyme inhibitors, we looked at use in patients for whom they have been proven useful; for calcium channel blockers, we looked at use in patients for whom they have been shown to be hazardous.
Medication use was recorded on the case report form both at randomization and at discharge. Although aspirin use was recommended in the protocol, the use of all other drugs at discharge was left to the discretion of the primary physician. Calcium channel blockers were divided into two categories on the case report form: nifedipine, diltiazem, or verapamil; or felodipine or amlodipine. We included only those in the former category in our analysis, as more evidence exists supporting their potential hazard in our population (7–13). Based on a random sample review of 10% of all case report form variables, there was >98% agreement between discharge medication information found on the case report form and that found in the corresponding medical records.
Baseline characteristics of patients and enrolling hospitals were expressed as percentages. Medication use at discharge in all patients, potential patients and “ideal” candidates was expressed as a percentage of the overall population. Medication use was also determined for “ideal” patients in the U.S., Canada and Europe. We tested for significant differences in medication use by geographic region using chi-square statistics. Odd ratios and 95% confidence intervals (CIs) for the likelihood of an “ideal” patient receiving a drug at discharge were determined by gender, age category and previous hypertension or diabetes status.
One-year mortality for unstable angina patients was estimated at 12%, and mortality for unstable angina patients with reduced left ventricular function (ejection fraction <40% or congestive heart failure) was estimated at 21% (1–3,14,15). From these estimates, the expected number of deaths in 1,000 patients at 1 year is 120 for unstable angina patients overall and 210 for unstable angina patients with reduced left ventricular function. From randomized trial data, we know that aspirin reduces the incidence of reinfarction and mortality at 1 year by about 50% in patients with unstable angina (1,16,17), and beta-blockers afford a relative reduction in 1-year mortality in this population of about 21% (18–22). Angiotensin-converting enzyme inhibitors reduce mortality in patients with reduced left ventricular function by about 23% (14,15,23,24). The use of calcium channel blockers in patients with unstable angina has been associated with a 23% excess mortality for those with ischemic disease and poor left ventricular function (7–12). We then estimated the lives that could have been saved with 100% guideline adherence, using estimates of the relative mortality reduction afforded by each medication applied to the estimated mortality in the population treated (120 potential lives saved/1,000 patients treated for aspirin and beta-blockers and 210 potential lives saved/1,000 patients treated for the use of angiotensin-converting enzyme inhibitors and the avoidance of calcium channel blockers). We then adjusted the potential lives saved for the current rate of use of each drug in our population to determine the actual lives saved. The additional lives that would have been saved (if medication use followed guideline recommendations) reflected the difference between the potential lives saved and the actual lives saved with the current rate of use. The actual 1-year mortality in our population (6.8%; 641/7,743) was used to calculate the mortality reduction with ideal use by subtracting the calculated additional lives saved from actual deaths. This mortality reduction along with 95% CI is shown.
The baseline characteristics of the 7,743 patients with non–ST elevation acute coronary syndromes included in our analysis are shown in Table 1, as are the in-hospital complications and angiographic data. This population had a high incidence of previous coronary artery disease; 76% had prior angina, 31% had prior infarction and 12% had prior bypass surgery. Of this population without ST segment elevation at presentation, 41.6% later were classified as having had an enrollment myocardial infarction. Of the 4,392 who underwent angiography during hospitalization, only 4.4% were found to have no coronary artery disease (<25% stenosis in any coronary vessel), and 62% were found to have two- or three-vessel disease (≥50% stenosis).
Using the American Hospital Association files from 1994, we compared demographics and certifications of the 181 U.S. hospitals participating in GUSTO-IIb with the 5,123 other U.S. hospitals of the same service code. Study hospitals were larger and more likely to be accredited by the Joint Commission on Accreditation of Health Care Organizations, be affiliated with medical schools, have Accreditation Council for Graduate Medical Education training programs and have cardiac catheterization and surgery facilities than the average U.S. hospital of similar service type (Table 2).
Overall use of discharge medications
Overall use of discharge medications is shown in Table 3. Use is also shown for potential and “ideal” patients for each drug. For aspirin and beta blockers, overall use and potential use are identical, because all patients were considered potential patients for these therapies. The use of aspirin in “ideal” patients was only slightly higher than its use in all patients (85.6% vs. 83.2%). Although the use of beta-blockers in “ideal” patients was greater than in potential patients, still only 59.1% of “ideal” patients received this drug at discharge. The use of angiotensin-converting enzyme inhibitors was more than doubled in “ideal” patients compared with the overall population (51.7% vs. 22.5%). Likewise, the use of calcium channel blockers was nearly halved in patients who were selected as “ideal” for avoidance of this medication compared with the overall population (26.7% vs. 41.6%).
International use of discharge medications
The ideal use by international region of enrollment was similar to the ideal use in the overall population (Table 4). Canada differed significantly from the U.S. in its higher use of beta-blockers in “ideal” patients but not in the use of other medications. Europe differed significantly from the U.S. in the higher use of aspirin in “ideal” patients but not in the use of other medications. The use of angiotensin-converting enzyme inhibitors and the nonuse of calcium channel blockers in “ideal” patients did not differ significantly among regions.
Likelihood ratios for receiving medications at discharge
For each group of “ideal” patients, we compared the likelihood of receiving a drug at discharge by certain baseline characteristics (Table 5). Women and older patients in the “ideal” category were less likely to receive aspirin at discharge. For example, men were 50% more likely to receive aspirin at hospital discharge than women, and patients younger than age 65 were 25% more likely to receive aspirin than those over age 75. Older patients were more likely to receive beta-blockers when “ideal,” whereas younger patients were more likely to receive calcium channel blockers even though they were “ideal” for their avoidance. In addition, patients with hypertension were more likely to receive beta-blockers, and diabetics were more likely to receive calcium channel blockers, at discharge.
Mortality reductions with evidence-based use of medications
If medication use in “ideal” patients increased from its current rate to 100%, the number of additional lives that could be saved per 1,000 patients treated is shown in the last column of Table 6. We converted these numbers to the additional number of lives saved in our population of ideal patients if all ideal patients had been treated. The nine additional lives per 1,000 ideal patients treated that could be saved by increasing the use of aspirin from 85.6% to 100% translate into 56 additional lives in our population of 6,175 patients who are ideal for aspirin. Increased beta-blocker use could have resulted in an additional 71 lives saved of 6,467 “ideal” patients treated, and increased angiotensin-converting enzyme inhibitor use could have saved an additional nine lives of 433 “ideal” patients treated. The decreased use of calcium channel blockers could have saved an additional five lives of 393 “ideal” patients with their nonuse. In total, 141 lives could have been saved in our population of 7,743 patients, or 18 additional lives saved per 1,000 treated, with guideline-recommended medication use. The actual 1-year mortality in our population after hospital discharge was 8.3% (641 deaths); altering medication use toward the ideal could have reduced the 1-year mortality rate to 6.5% (500 deaths). This 6.5% mortality (95% CI 5.9% to 7.0%) translates into a relative reduction in mortality of 22% (95% CI 16% to 44%) with more evidence-based clinical practices. If medication use improved only 75% toward guideline recommendations, it would result in 106 lives saved in our population of 7,743 patients, yielding a 1-year mortality of 6.9% (535 deaths) and a relative risk reduction for mortality of 17%.
This study describes medication use in a large, international population of patients with non–ST elevation acute coronary syndromes, and is unique in estimating the impact on mortality that suboptimal adherence to guideline recommendations may have in this population. We found that the use of cardiac medications in patients with non–ST elevation acute coronary syndromes falls short of guideline recommendations across international practice, and that significant mortality reductions may be achieved with a more evidence-based approach to the use of aspirin, beta-blockers, angiotensin-converting enzyme inhibitors and calcium channel blockers.
The purpose of our analysis was to determine the adherence to these medical recommendations in general practice by applying published guidelines in broad strokes to a large population. Therefore, we used purposefully narrow criteria to select “ideal” patients, and may have excluded some patients from this category who might have benefited from these therapies. For example, we excluded patients with reduced left ventricular function from consideration for beta-blockers, although recent studies have indicated the potential for mortality reduction with these agents (25). In addition, we excluded felodipine and amlodipine from our analysis of calcium channel blockers, as these agents have not been shown to be hazardous and may be useful as adjunctive therapy in patients with ischemic heart disease (13). With these caveats in mind, our analysis provides a broad view of guideline-recommended uses of each medication and does not specify the ideal medical regimen for each patient.
Although guideline-recommended use of medications has been well described in patients with acute myocardial infarction, less is known about their “ideal” use in an unstable angina population. Although the heterogeneity in clinical presentation and acuity of an unstable angina population can make applying uniform guidelines challenging, our population was uniquely suited for this analysis due to a high degree of coronary disease severity and acuity. For example, many had prior cardiac diagnoses, 42% were diagnosed with an enrollment infarction and 84% of those undergoing angiography had significant coronary disease. A previous analysis of the unstable angina population in GUSTO-IIb has shown 97.6% of these patients to be in the intermediate and high risk categories according to unstable angina guidelines (26). In addition to the high risk and likelihood of coronary disease of our patients, the GUSTO-IIb enrolling hospitals were high volume, cardiovascular referral centers from around the world. Based on these patient and hospital characteristics, we expected a higher than average adherence to guideline recommendations in our population compared with general practice, even though we have shown this adherence to be suboptimal.
The Thrombolysis in Myocardial Infarction trial (TIMI)-III Registry Study Group recently reported medication use without adjustment for indications and contraindications in a similar population of patients with unstable angina and non–Q wave myocardial infarction (27). The overall rates of use in their U.S. population were 82.5% for aspirin, 39.5% for beta-blockers and 54.3% for calcium channel blockers. Comparing the overall rates in their registry with those in our analysis, we found a similar use of aspirin, a lower use of calcium channel blockers and a higher use of beta-blockers. Our study shows, however, that indications and contraindications do not explain the gap between actual and recommended use of medications seen in large cardiovascular trial registries.
Compared with reports of the medication use in the acute myocardial infarction and congestive heart failure populations, we found similar patterns of underuse of guideline-recommended medical therapies in our population (5,21,28,29). The 86% use of aspirin in our unstable angina population compares favorably with reported rates of aspirin use after ST elevation myocardial infarction, which have ranged from 81% to 84% (5,28,30). Since aspirin use was recommended in the protocol, the rate of aspirin use in our population may be higher than that found in general practice. Reports of in-hospital use of beta-blockers after ST elevation myocardial infarction range from 45% to 53% (5,22,28,30), compared with the 59% overall use of beta-blockers in our analysis. The use of angiotensin-converting enzyme inhibitors in patients admitted for management of heart failure was 47% in one study (29), and angiotensin-converting enzyme inhibitor use in patients with reduced left ventricular function after infarction has been reported to be <40% (31). We report a higher, 51% use in patients admitted for unstable angina with a low ejection fraction. There are no other reports with which to compare the use of calcium channel blockers in patients for whom their use is contraindicated. Overall, our findings in a large population of patients with non–ST elevation acute coronary syndromes are similar to those in registries of patients with ST elevation acute myocardial infarction.
There were minimal differences in the use of discharge medications internationally, with two exceptions. One was the higher use of beta-blockers in Canada, also found in the TIMI-III Registry (27), and the other was the higher use of aspirin in Europe. Despite the slightly better rates in Canada and Europe, use of medications consistently fell short of recommendations regardless of place of enrollment. Therefore, improvements in guideline-recommended use of medications are needed across international practice.
Interestingly, certain patient characteristics influenced the likelihood of “ideal” patients receiving medications at discharge. We found that gender and age influenced the likelihood of receiving aspirin, with women and older patients being less likely to receive aspirin at discharge. In addition, younger patients and diabetics were both more likely to be given calcium channel blockers and less likely to be given beta-blockers. These differences may result from a perception that certain groups are less likely to have coronary artery disease or from concerns over side effect profiles of these agents in certain patients. Awareness of these trends will help physicians focus their efforts to improve guideline adherence toward specific subgroups of patients in whom medications are now underused.
Estimate of benefit
Observational reports suggest that patients who are “ideal” for medications and actually receive them have better survival than those who do not (28,32,33). However, any nonrandomized comparison of survival by discharge medication is obviously confounded by unknown factors in the selection of patients for treatment. Patients not receiving therapy are likely to be at higher risk due to the presence of contraindications and therefore may have a higher mortality. In addition, since the absolute benefit of a treatment depends on the inherent risk of the patient subgroup, the absolute benefit may be even larger among this untreated group of patients. Because of these factors, we elected to predict mortality reductions using literature estimates of benefit instead of comparing observed 6-month and 1-year mortality for “ideal” patients within our population. In addition, the literature estimates of baseline mortality are unaffected by current rates of medication use in our population. By our methods, we estimated that by increasing medication use to 100% of the ideal, 141 lives out of our population of 7,743 could have been saved at 1 year for a 22% mortality reduction. This translates to 18 lives saved per 1,000 patients treated per year in a more evidence-based clinical practice. Assuming there are undetected but legitimate reasons for underuse of beneficial therapies, we also calculated the mortality reductions for moving only 75% toward the goal of 100% ideal and found that a 17% relative mortality reduction would still result. This rivals and surpasses the mortality benefits seen with many of the newer therapies for acute coronary syndromes. For comparison, treatment with hirudin in the GUSTO-IIb trial afforded a 9% relative risk reduction in death and myocardial infarction at 30 days compared with heparin (4).
Our study has several limitations. First, the AHCPR guidelines are published in the U.S., and non-U.S. investigators may be less aware of their content. They do, however, contain practice recommendations that are based on the results of widely known international, randomized clinical trials. Second, U.S. hospitals participating in GUSTO-IIb differ in many service indicators from national averages in that they are larger and more often academically affiliated, and they provide more cardiovascular services. In addition, our non-U.S. hospitals also represent the large, tertiary care referral centers within their geographic regions. Therefore, practice patterns in these hospitals may not accurately represent medical practice in the community. Third, information that was unavailable through the case report form may have played a crucial role in the decision to prescribe or withhold a medication. Patient-reported aspirin intolerance may have resulted in the use of antiplatelet agents other than aspirin. In addition, limitations in blood pressure may have prevented the use of more than one medication before discharge, although there may have been several for which the patient was “ideal.” The frequent use of calcium channel blockers in our population may be due partly to a hesitance on the part of hospital physicians to alter preexisting, presumably effective outpatient medical regimens. In addition, we determined medication use only at discharge. We did not account for drugs used earlier during hospitalization that were discontinued for intolerance, or those whose use was planned to begin after discharge. Due to these limitations, we suspect that actual use may be closer to appropriate than our numbers would suggest. We do not have information on the dosages of prescribed drugs, and therefore we do not know whether patients received the same dosages proven to reduce mortality in clinical trials. Finally, our literature-based prediction of mortality reduction was calculated by adding the individual effect of each medication, when in practice, the combined medical regimen may affect clinical outcome in a manner that is not purely additive.
There is significant room for improvement in clinical practice by increasing the use of medications known to improve outcomes in patients with non–ST elevation acute coronary syndromes. Too often, patients who are “ideal” for these therapies are discharged without them. This may be a result of inattention to the medical regimen at discharge, to lack of knowledge about guideline recommendations or to physicians’ concern over adverse effects of these drugs in certain subgroups. Although guidelines are not a mandate for compliance with a medical regimen, if they were followed more closely, significant mortality reductions could be expected in patients with unstable angina worldwide. Thus, as the medical community develops new therapies for treatment of acute coronary syndromes, physicians must remember to use the basic medications proven over time to be of benefit in these patients.
We would like to acknowledge the excellent editorial assistance of Ms. Patricia Williams in the preparation of the manuscript.
Selection criteria for “ideal” patients
Aspirin use at discharge:
• Inclusion criteria (n = 7,743): All patients.
• Exclusion criteria (n = 1,568): Discharged on warfarin (34%), major bleeding (34%), platelet count <100,000/mm3(5%) or hemoglobin <10 mg/dL (61%).
• “Ideal” patients (n = 6,175) = potential patients (n = 7,743) − excluded patients (n = 1,568).
Beta-blocker use at discharge:
• Inclusion criteria (n = 7,743): All.
• Exclusion criteria (n = 1,276): No significant angiographic coronary disease (29.5%), ejection fraction <40% (26.5%), in-hospital congestive heart failure (16.2%), pulmonary edema (20.1%), shock (4.5%), second- or third-degree heart block (8.5%), chronic obstructive pulmonary disease (17.0%).
• “Ideal” patients (n = 6,467) = potential patients (n = 7,743) − excluded patients (n = 1,276).
Angiotensin-converting enzyme inhibitor use, patients with reduced ejection fraction at discharge:
• Inclusion criteria (n = 513): Ejection fraction <40% (66%), in-hospital congestive heart failure (40%).
• Exclusion criteria (n = 2,708): History of chronic renal insufficiency (4%), creatinine >177 μmol/L (2.0 mg/dL, 4%), ejection fraction >40% (95%).
• “Ideal” patients (n = 433) = potential patients (n = 513) − excluded patients (n = 80).
No calcium-channel blocker use, patients with reduced ejection fraction at discharge:
• Inclusion criteria (n = 513): Discharged with ejection fraction <40% (66%), in-hospital congestive heart failure (40%).
• Exclusion criteria (n = 2,870): Ejection fraction >40% (90%), atrial fibrillation (17%).
• “Ideal” patients (n = 393) = potential patients (n = 513) − excluded patients (n = 120).
☆ This study was supported by Ciba-Geigy Corporation (Summit, New Jersey), and Advanced Cardiovascular Systems (Mountain View, California).
- Agency for Health Care Policy and Research
- confidence interval
- Global Use of Strategies to Open Occluded Arteries in Acute Coronary Syndromes
- Thrombolysis in Myocardial Infarction trial
- Received February 18, 1998.
- Revision received July 23, 1998.
- Accepted August 6, 1998.
- American College of Cardiology
- Braunwald E,
- Mark D,
- Jones R,
- et al.
- The Chest Pain Study Group,
- White L.D,
- Lee T.H,
- Cook E.F,
- et al.
- Ryan T.J,
- Anderson J.L,
- Antman E.M,
- et al.
- Muller J.E,
- Turi Z.G,
- Pearle D.I,
- et al.
- Furberg C.D,
- Psaty B.M,
- Meyer J.V
- Held P,
- Yusuf S,
- Furberg C.D
- The Adverse Experience Committee and the Multicenter Diltiazem Postinfarction Research Group,
- Goldstein R.E,
- Boccuzzi S.J,
- Cruess D,
- Nattel S
- Brand D.A,
- Newcomer L.N,
- Freiburger A,
- Tian H
- Heidenreich P.A,
- Lee T.T,
- Massie B.M
- Alexander KP, Peterson ED, Hellkamp AS, Granger CB, Topol EJ, Califf RM, for the GUSTO IIb Investigators
- Rogers WJ, Bowlby LJ, Chandra NC, et al. for the Participants in the National Registry of Myocardial Infarction
- Koenig W,
- Lowel H,
- Lewis M,
- Horman A