Author + information
- Received April 10, 1998
- Revision received May 4, 1998
- Accepted May 15, 1998
- Published online September 1, 1998.
- Peter A. McCullough, MD, MPH∗,* ( )(, )
- William W. O’Neill, MD, FACC∗,
- Mariann Graham, BSN†,
- Robert J. Stomel, DO, FACC‡,
- Felix Rogers, DO, FACC§,
- Shukri David, MD, FACC∥,
- Ali Farhat, MD∗,
- Rasa Kazlauskaite, MD†,
- Majid Al-Zagoum, MD∗ and
- Cindy L. Grines, MD, FACC†
- ↵*Address for correspondence: Dr. Peter A. McCullough, Cardiovascular Division, Henry Ford Hospital, Henry Ford Health System, Center for Clinical Effectiveness, One Ford Place, Suite 3C, Detroit, Michigan 48202 or; worldwide web: http://www.hfhs-cce.org/
Objectives. The purpose of this study was to determine if early triage angiography with revascularization, if indicated, favorably affects clinical outcomes in patients with suspected acute myocardial infarction who are ineligible for thrombolysis.
Background. The majority of patients with acute myocardial infarction and other acute coronary syndromes are considered ineligible for thrombolysis and therefore are not afforded the opportunity for early reperfusion.
Methods. This multicenter, prospective, randomized trial evaluated in a controlled fashion the outcomes following triage angiography in acute coronary syndromes ineligible for thrombolytic therapy. Eligible patients (n = 201) with <24 h of symptoms were randomized to early triage angiography and subsequent therapies based on the angiogram versus conventional medical therapy consisting of aspirin, intravenous heparin, nitroglycerin, beta-blockers, and analgesics.
Results. In the triage angiography group, 109 patients underwent early angiography and 64 (58%) received revascularization, whereas in the conservative group, 54 (60%) subsequently underwent nonprotocol angiography in response to recurrent ischemia and 33 (37%) received revascularization (p = 0.004). The mean time to revascularization was 27 ± 32 versus 88 ± 98 h (p = 0.0001) and the primary endpoint of recurrent ischemic events or death occurred in 14 (13%) versus 31 (34%) of the triage angiography and conservative groups, respectively (45% risk reduction, 95% CI 27–59%, p = 0.0002). There were no differences between the groups with respect to initial hospital costs or length of stay. Long-term follow-up at a median of 21 months revealed no significant differences in the endpoints of late revascularization, recurrent myocardial infarction, or all-cause mortality.
Conclusions. Early triage angiography in patients with acute coronary syndromes who are not eligible for thrombolytics reduced the composite of recurrent ischemic events or death and shortened the time to definitive revascularization during the index hospitalization. Despite more frequent early revascularization after triage angiography, we found no long-term benefit in cardiac outcomes compared with conservative medical therapy with revascularization prompted by recurrent ischemia.
In the past decade, advances in terms of mortality reduction have been made in the treatment of acute transmural myocardial infarction with aspirin, beta-blockers and reperfusion therapy with thrombolytics or primary angioplasty (1–13). Unfortunately, similar inroads have not been made with respect to other acute coronary syndromes, including nontransmural infarctions and transmural infarctions that are not eligible for reperfusion therapy (14). In fact, several lines of evidence indicate thrombolytic therapy is ineffective or potentially harmful in patients without classic electrocardiographic (ECG) findings of transmural infarction (15–25). Retrospective reviews have shown the in-hospital mortality for those who are not eligible for thrombolysis or reperfusion to be twofold greater over those eligible and treated with reperfusion therapy (26,27). Recent studies have indicated that patients with acute coronary syndromes and evidence of myonecrosis from sensitive enzymatic markers without elevations of creatine kinase (CK) have a conferred increased risk of death over the next 30 d irrespective of revascularization status (28–32). These findings imply that preempting ischemic events rather than acting after they have occurred may be important in reducing morbidity and mortality in those patients who are not candidates for reperfusion therapy. Our study was designed to define the role of early, catheterization-based therapy in patients with acute coronary syndromes who are deemed likely to be suffering from acute myocardial infarction (AMI), but are ineligible for thrombolysis based on either the lack of classic ECG changes or the presence of contraindications to thrombolytics. We hypothesized that triage angiography with early revascularization, when indicated, could reduce recurrent ischemic events and death.
This study was approved by the Human Investigations Committees at the four participating Southeast Michigan community hospitals and subjects gave informed consent. William Beaumont Hospital in Royal Oak and Providence Hospital in Southfield provided 24-h angiography and angioplasty service with bypass surgery programs. Botsford Hospital in Farmington Hills, and Riverside Hospital in Trenton, offered angiography during working hours and referred patients to the larger centers for off-hours angiography, angioplasty and bypass surgery. Subjects were those patients 18 years and older who presented to the emergency department with an acute chest pain syndrome consistent with AMI (high clinical suspicion for AMI with or without immediate enzymatic confirmation) and who were considered ineligible for thrombolysis because of a lack of diagnostic ECG changes, symptoms lasting longer than 6 h or because of increased bleeding or stroke risks. None of these inclusion criteria mandated entry into the trial. They were designed to provide clinicians with categories to characterize patients during the clinical judgment for thrombolytic eligibility. The inclusion criteria allowed for a spectrum of ECG changes, as illustrated in Table 1. Exclusion criteria were symptoms lasting for more than 24 h or an absolute indication or contraindication to cardiac catheterization.
Randomization and treatment
After informed consent was obtained, patients were randomized according to a computer-generated scheme and cards placed in sealed, ordered envelopes. Protocol guidelines called for all patients to be treated with 325 mg of chewed aspirin, heparin 5,000–10,000 U intravenous (IV) bolus, followed by a 10 U/kg/h infusion, IV nitroglycerin 10–30 μg/min initially and then titrated, and IV metoprolol up to 15 mg or IV propranolol up to 8 mg as tolerated to achieve a heart rate < 70. Subjects randomized to the conservative arm were then admitted to a monitored bed and received continued medical therapy and noninvasive evaluation encouraged by the protocol. There were no mandated noninvasive tests or schedules in the conservative arm. Criteria for crossover to a nonprotocol catheterization in the first 24 h were persistent pain despite maximal medical therapy >4 h in duration, recurrent pain with ECG changes once pain free or hemodynamic instability. Beyond the 24-h window, subjects underwent catheterization based on ischemic events or physician preference, which was representative of the standard care at the individual institution. Subjects randomized to triage angiography were taken as soon as possible directly to the catheterization laboratory from the emergency department. All triage angiography patients underwent catheterization within 24 h of arrival to the hospital. The protocol encouraged decision making based on the angiogram and discouraged additional noninvasive testing. Subjects with an identifiable culprit lesion and suitable anatomy underwent mechanical revascularization with percutaneous transluminal coronary angioplasty (PTCA), atherectomy or stenting as decided by the operator. All cases referred for coronary artery bypass graft surgery (CABG) had the operation performed on the index hospitalization. Early discharge was encouraged in both groups.
Long-term follow-up was carried out by trained telephone interviewers blinded to the randomization arm. Subjects underwent a structured questionnaire at a median time of 21 months (12 months minimum and 52 months maximum) after the index hospitalization. Twelve subjects, unable to be contacted, were submitted to the National Center for Health Statistics National Death Index for vital status at a minimum of 1 year after the index event.
Intention-to-treat principles were used. Univariate statistics were reported as proportions or means and comparisons were made with chi-square, Fisher’s exact test, Kruskal-Wallis H-test and analysis of variance (ANOVA) as appropriate. The primary composite endpoint of recurrent ischemic events or death was evaluated with absolute proportions and the calculated percent risk reduction with 95% exact mid-p confidence intervals. The number needed to treat was calculated as the reciprocal of the absolute risk reduction (33). Secondary endpoints including length of stay (LOS) and hospital costs were compared using ANOVA, chi-square for proportions and Fisher’s exact test as appropriate. Outliers for both LOS and cost were excluded if they were more than 2 standard deviations from the mean of the parent distribution. True hospital costs, given in 1995 dollars, were determined from internal accounting programs at William Beaumont Hospital and from the hospital Medicare cost-to-charge ratio at the other three hospitals. Reliable physician cost data were not available. For patients who had stays at two hospitals, LOS and costs were combined from the two facilities. The sample size calculation of 225 in each arm was based on an expected rate of the composite endpoint of 40%, an effect size of 30% (absolute reduction), 1-β of 0.80 and α of 0.05 two-tailed. At 201 patients, one interim analysis was performed with the randomization being unblocked, yielding 111 in the triage angiography arm and 90 in the conservative arm. Because multiple hospitals were used, and each had sets of randomization cards in sealed envelopes that were not equal at set enrollment points, the study ended with unequal numbers in each group. Statistical significance was chosen at the α < 0.05 level. Survival analysis was performed using Kaplan-Meier curves and the log rank test for differences between the randomization arms.
Variables and definitions
Acute myocardial infarction was defined as a characteristic rise in total CK and myocardial bands (MB) in the serum with the peak CK > 3.8 μkat/L (230 U/L) in males, and CK > 2.5 μkat/L (150 U/L) in females, and an MB index (MB fraction ng/ml/total CK U/L × 100) > 3% in both sexes. Recurrent ischemia was defined as chest discomfort with or without ECG or hemodynamic changes, or reinfarction. Angioplasty success was defined as <50% residual lesion stenosis and thrombolysis in myocardial infarction (TIMI) grade III flow in the culprit vessel. These outcomes were determined prospectively, during the hospital stay, by an independent research nurse and all ECGs were over-read by a blinded physician (P.A.M.). Stroke was defined as a confirmed complication by a neurologic consultant’s clinical evaluation. Transient azotemia was considered as a rise in serum creatinine >25% over the baseline value. A bleeding complication was defined as any bleeding event that required at least one transfusion of red blood cells. A vascular complication was defined as any catheter-related vascular injury requiring vascular surgical repair. Revascularization included any type of percutaneous intervention such as PTCA, atherectomy or stenting and CABG.
Baseline characteristics for the randomization arms are given in Table 1. The randomization held for all baseline characteristics except for a history of non–insulin-dependent diabetes mellitus (NIDDM) (9% vs. 22%) and a history of previous smoking (4% vs. 19%), both of which were more prevalent in the conservative group (p = 0.009 and p = 0.0004, respectively). There were no significant differences between the groups with respect to presentation features or reasons for reperfusion ineligibility. The mean time from symptom onset to study entry was 9 ± 7 h versus 9 ± 6 h, p = 0.53, for the triage angiography and conservative groups, respectively. Of note, 36 (22%) and 25 (28%) of the triage angiography and conservative groups, respectively, presented with ST segment elevation, which was persistent in 14 (13%) and eight (9%). The distribution of cases by time to emergency department presentation in the ST elevation subgroup was as follows: within 6 h of symptoms, 24 versus 14; between 6 and 12 h, three versus six; between 12 and 24 h, seven versus five; and greater than or equal to 24 h, two versus zero, for the triage angiography and conservative groups, respectively. Irrespective of the time to presentation, the patients with ST segment elevation had an average of 1.8 contraindications to thrombolysis. This compares with the whole study population, which had an average of 1.6 contraindications to thrombolysis.
Medical and interventional treatment
Proportions and types of medical, interventional and surgical therapy received by the patient groups are depicted in Table 2. Patient flow and revascularization outcomes are depicted in Figure 1. Overall, 64 of 111 (58%) versus 33 of 90 (37%), p = 0.004, in the triage angiography and conservative groups, respectively, received revascularization therapy. Of those who underwent catheterization, each group had similar findings with respect to the proportion of patients revascularized after the angiogram 64 of 109 (59%) versus 33 of 54 (61%), p = 0.77. Of these 54 angiography cases in the conservative group, 27 (50%) underwent the procedure because of ischemia-driven events such as persistent or recurrent chest discomfort, hemodynamic complications or positive exercise treadmill test results. The remainder underwent the procedure as part of the physician’s or patient’s preference. There was no difference in the composite endpoint of in-hospital recurrent ischemic events or death in the physician-preference versus protocol or ischemia-driven angiography groups (4 of 27 [15%] vs. 29 of 136 [21%], p = 0.52).
Of the 163 patients who ultimately underwent angiography, the triage angiography and conservative groups were similar with respect to the extent of disease and to the location of the culprit lesion with the exception of the circumflex system (Table 3). There were less circumflex or obtuse marginal culprit lesions in the triage angiography arm (10 [9%] vs. 11 [20%], p = 0.05). There were no differences between the mean percent stenosis of the culprit lesion or the number of occluded culprits between the randomization groups, with overall 33 of 163 cases (20%) having an occluded vessel at the time of angiography.
Myocardial infarction, in-hospital recurrent ischemic events and death
Acute myocardial infarction was confirmed in 57 (51%) and 49 (54%) of the triage angiography and conservative groups, respectively (p = 0.81) (Table 4). Of those with AMI, there were no differences between peak CK values (16 ± 22 μkat/L vs. 14 ± 16 μkat/L and 962 ± 1,339 U/L vs. 835 ± 963 U/L, p = 0.47). Suspected acute MI was used as an inclusion criteria into the trial and the initial rise in CK was not included in any study endpoint. There were no differences between the groups with respect to reinfarction or death (3 [3%] vs. 3 [3%], p = 0.5). However, the expanded aggregate of chest pain with ECG or hemodynamic changes, reinfarction or death occurred in 5 (5%) and 12 (13%) of the triage angiography and conservative groups, respectively (40% risk reduction, 95% CI 15–58%, p = 0.03). Similarly, the composite endpoint of all recurrent ischemic events or death occurred in 14 (13%) and 31 (34%), yielding a 45% risk reduction (95% CI 27–59%, p = 0.0002), and the number needed to treat of five patients to prevent one composite endpoint.
Hospital length of stay and costs
Economic endpoints of LOS and cost are given in Table 4. There were no differences in mean LOS between the groups after adjustment for outliers. However, 39 (35%) versus 15 (17%) of the triage angiography and conservative groups, respectively, were discharged in 2 or less days (p = 0.003). Similarly, of those who underwent PTCA, 36 of 48 (75%) versus 13 of 27 (48%) were discharged in 5 or less days (p = 0.02); and of those who underwent CABG, 12 of 18 (67%) versus 1 of 7 (14%) were discharged in 10 or less days (p = 0.03).
A comparison of mean hospital costs in 1995 dollars is given in Table 4. The overall cost for all 201 patients had the following distribution characteristics: mean = $12,071 ± $24,057, range $1,868–$303,659 and median = $7,286. There were no significant cost differences between the randomization groups stratified by procedural subgroups (Table 4).
Stroke and vascular complications
There were no significant differences between the triage angiography and conservative groups in terms of stroke: 2 (2%) versus 2 (2%) (p = 0.77). Each group had one hemorrhagic and one nonhemorrhagic stroke. Transient azotemia occurred in 2 (2%) versus 3 (3%) (p = 0.81). Renal failure requiring dialysis did not occur. Vascular complications requiring surgical repair occurred in 1 (1%) versus 0 (0%) (p = 0.55) (Table 2). There was a trend for more transfusions in the triage angiography arm (11 [10%] vs. 4 [5%], p = 0.17), however, the majority (12 of 15) of these transfusions occurred in the postoperative period after CABG.
Time from admission to revascularization
Times from symptom onset to emergency department arrival, angiography, PTCA and CABG are given in Table 2. The mean time from admission to any form of revascularization was 27 ± 32 versus 88 ± 98 h for the triage angiography and conservative groups, respectively (p = 0.0001). Similarly, the mean waiting times to PTCA or CABG were shorter in the triage angiography group (Fig. 2).
Late cardiac events
The median time from the index event to follow-up was 21 months, with a minimum of 12 and a maximum of 52 months. There were no significant differences between the randomization arms with respect to repeat hospitalizations, subsequent angiography, late PTCA or late CABG (Table 4, Fig. 3). Likewise, there were no differences in the overall cumulative (from the time of initial presentation to long-term follow-up) incidences of PTCA, CABG, recurrent nonfatal AMI or all-cause death (Fig. 4). The composite endpoint of cumulative nonfatal MI or death occurred in 15 (14%) versus 11 (12%) in the triage angiography and conservative groups, p = 0.6. Similarly, the Kaplan-Meier estimates revealed no differences between the groups over the follow-up period with respect to this endpoint (p = 0.8) by the log rank statistic (Fig. 5).
Bypass surgery mortality
Deaths after bypass surgery occurred on the index hospitalization in 1/18 (6%) and 1/6 (17%) (p = 0.4) in the triage angiography and conservative groups, respectively. Mortality after late CABG procedures performed during the follow-up period was 0 of 2 (0%) and 1 of 3 (33%), yielding an overall bypass surgery mortality rate of 1 of 20 (5%) and 2 of 9 (22%) (p = 0.22) in the two groups, respectively. Putting all cases together, bypass surgery mortality accounted for 3 of 21 (14%) of the deaths observed in this study.
Because of the high crossover rate (60%) in our trial, we performed an efficacy analysis with comparisons made with respect to patients who received early revascularization versus all others. Of the 201 subjects randomized, 97 (48%) received revascularization with PTCA or CABG on the index hospitalization and 13 (7%) received revascularization during the follow-up period (restenosis cases excluded), leaving 91 (45%) who were treated medically without revascularization. There were 21 versus 24 repeat hospitalizations, one versus three recurrent MIs and nine versus eight deaths, in the early revascularization group compared with all others. When the late events were adjusted for ischemia-guided revascularization, the comparisons above revealed 21 versus 11 hospitalizations (RR = 1.8, CI = 0.9–3.5, p = .08), one versus one recurrent MIs and nine versus eight deaths (all other p > 0.10) for the early revascularization versus medically treated groups (without revascularization). In addition, survival analysis by revascularization status showed no difference between early versus late or any versus none with the log rank test (all p > 0.10).
Studies of invasive versus conservative management in acute syndromes
Acute coronary syndromes ineligible for thrombolysis are a common cause for hospitalization and subsequent morbidity in terms of cardiac events in the United States (34). This trial is now one of three such studies to examine the issue of timing of angiography and revascularization. In the TIMI IIIB trial, early angiography occurred between 18 and 48 hours (reported mean of 1.5 d or 36 h) in 373 subjects versus 371 who received conservative care (nonthrombolytic arms of the factorial) (20). In the partially reported Veterans Administration Non-Q-Wave Strategies In-Hospital (VANQWISH) trial, early angiography was carried out at a median of 48 h after hospital admission in 462 versus 458 who received ischemia-guided care (35–37). Our trial is the first to employ angiography as a triage mechanism where the procedure was performed as early as possible to the presentation of symptoms and entry into the medical system, via community hospital or tertiary center.
The TIMI IIIB, VANQWISH and MATE trials all exhibited the difficulty in adhering to a conservative protocol arm, with in-hospital crossover to angiography rates of 67%, 24% and 60% respectively (20,21,35–37). The VANQWISH trial had the best protocol adherence, probably attributable to the fact that subjects were allowed to be enrolled after 24 h and represented a more stable group with completed non-q-wave infarctions (35–37). In fact, the VANQWISH protocol required subjects to be enrolled after 24 h and have no evidence of recurrent ischemia or hemodynamic compromise (38). All three trials demonstrated the principle that mandated angiography will lead to earlier revascularization by PTCA or CABG. In addition, the MATE trial demonstrated the concept that reducing the time from the index event to angiography limits the opportunity for recurrent ischemic events to occur. This reduction in recurrent ischemia, however, did not appear to bear on the long-term result of all-cause mortality in the treatment groups.
Invasive strategies and cardiac enzymes
Our finding of overall 20% of culprit vessels being occluded at the time of angiography (33 of 163 of cases where angiography was performed) is consistent with the work of Kulick and Rahimtoola, which suggests that if reperfusion could occur early enough, a benefit may be observed (39). Furthermore, analyses from multiple trials of acute coronary syndromes indicate that even small infarctions detected by troponin testing and not by traditional CK measurements indicate a poorer short-term prognosis (28–32). These observations led to the notion that early revascularization would be beneficial in such patients. Troponin testing was not used in the early angiography decision in the TIMI IIIB, VANQWISH or the MATE study, and perhaps this or some other triage tool should be tested next in a randomized trial to determine which strategy, if any, is superior to the conventional, ischemia-guided treatment of acute ischemic syndromes (20,21,35–37).
Clinical effectiveness of early angiography
The immediate clinical utility of coronary angiography has probably been understated in all three trials that led to a focus on long-term mortality as the ultimate endpoint. The advantage of the angiogram over conventional triage tools is the decision-enhancing capability provided to the physician. Knowledge of the coronary anatomy, left ventricular function and filling pressures permits triage of patients to medical therapy, immediate percutaneous intervention or bypass surgery. Furthermore, the triage angiogram helps provide definitive disposition of discharge to home, telemetry unit or coronary care unit. These benefits have been difficult to quantitate in previous studies of acute coronary syndromes management (39,40). The shortened time to revascularization and reduction in recurrent ischemic events both indirectly reflect the enhanced clinical efficiency of the triage angiography strategy.
Triage angiography benefited those patients encaptured in our study who had noncoronary chest discomfort syndromes (proved angiographically in 13%). These patients accounted for 44% of those patients able to be discharged from the hospital in 2 d or less. Conversely, only 17% of the conservative group patients were able to be discharged to home in less than 2 d. This was attributable to the common delays of “rule-out AMI” enzymatic protocols, scheduling of noninvasive testing and physician review of the data (41). In addition, triage angiography appeared to be safe in that no increases in stroke, vascular or renal complications were observed.
Limitations of our study, common to studies of this nature, include the impossibility of blindedness, the interim analysis performed at an “unblocked” level of the randomization and the high utilization of angiography in the conservative arm. The former two could have favored the triage angiography group, whereas the latter clearly favored the conservative group in terms of primary outcome assessment. The randomization held in terms of major confounders such as age, sex and previous history of coronary disease, and thus these factors are unlikely to have influenced the observed outcome. Our study, like many randomized trials, ultimately selected a relatively low-risk study group with an in-hospital mortality rate of 2% and a cumulative 10% mortality rate at 21 months of follow-up. These rates are considerably less than the retrospective series reported by Cragg and colleagues and Behar and coworkers, and reflect the inherently difficult inclusion of high-risk patients into clinical trials with invasive versus conservative randomizations (26,27). Finally, this study like others with logistically challenging randomization protocols, including TIMI IIIB and VANQWISH, was not powered to show a long-term recurrent myocardial infarction or mortality difference in treatment allocation (20,21,35–37). A study powered (80% power, α = 0.05) to show at least a 25% reduction in mortality would require approximately 8,000 subjects, far more than the combined 2,594 (1,865 not exposed to thrombolytics) available for analysis from TIMI IIIB, VANQWISH and MATE (20,21,35–37).
In conclusion, triage angiography applied to those patients with acute coronary syndromes considered to be ineligible for thrombolytic therapy provides a diagnostic and treatment platform that reduces the composition of recurrent ischemic events or in-hospital death, reduces the waiting time for definitive revascularization and is not associated with increased costs over conventional medical therapy and observation. These short-term benefits, however, do not translate into measurable long-term differences in the rates of rehospitalization, repeat angiography, late revascularization, recurrent myocardial infarction or death.
We are indebted to James A. Goldstein, MD, Gerald C. Timmis, MD, Gregg W. Stone, MD, and W. Douglas Weaver, MD. This group provided critical review and support to the trial. We would also like to express gratitude to Kristin B. McCabe, BA, who assisted in the final preparation of this manuscript.
The following centers and investigators contributed to patient recruitment in the MATE Trial: William Beaumont Hospital (P.A.M., M.G., C.L.G., W.W.O.), n = 168; Botsford Hospital (R.J.S.), n = 15; Riverside Hospital (F.R.), n = 10; and Providence Hospital (S.D.), n = 8.
Presented in parts at the 69th and 70th Scientific Sessions of the American Heart Association, November 13, 1996, New Orleans, LA, November 12, 1997, Orlando, FL, and at the 21st Annual Meeting of the Society for Cardiac Angiography and Interventions, May 13–16, 1998, Montreal, Canada.
- acute myocardial infarction
- coronary artery bypass graft surgery
- congestive heart failure
- creatine kinase
- length of hospital stay
- myocardial infarction
- non-insulin–dependent diabetes mellitus
- percutaneous transluminal coronary angioplasty
- Thrombolysis in myocardial infarction
- Received April 10, 1998.
- Revision received May 4, 1998.
- Accepted May 15, 1998.
- American College of Cardiology
- The MIAMI Trial Research Group
- Hjalmarson A.,
- Herlitz J.,
- Holmberg S.,
- et al.
- Gibbons R.J.,
- Holmes D.R.,
- Reeder G.S.,
- et al.
- Cohen M.,
- Xiong J.,
- Parry G.,
- et al.
- O’Neill W.W.,
- Weintraub R.,
- Grines C.L.,
- et al.
- Schreiber T.L.,
- Rizik D.,
- White C.,
- et al.
- Williams D.O.,
- Topol E.J.,
- Califf R.M.,
- et al.
- Nicklas J.M.,
- Topol E.J.,
- Kander N.,
- et al.
- Ambrose J.A.,
- Torre S.R.,
- Sharma S.K.,
- et al.
- The TIMI IIIB Investigators
- Anderson H.V.,
- Cannon C.P.,
- Stone P.H.,
- et al.
- White H.D.,
- French J.K.,
- Norris R.M.,
- et al.
- Ambrose J.A.,
- Almeida O.D.,
- Sharma S.K.,
- et al.
- Roberts M.J.,
- McNeill A.J.,
- Dalzell G.W.,
- et al.
- The Global Use of Strategies to Open Occluded Coronary Arteries (GUSTO) IIa Investigators
- Stubbs P.,
- Collinson P.,
- Moseley D.,
- Greenwood T.,
- Noble M.
- Cook R.J.,
- Sackett D.L.
- Braunwald E.,
- Jones R.H.,
- Mark D.B.,
- et al.
- Holland M.,
- Crawford M.H.,
- Ferry D.R.,
- Pepine C.J.,
- Boden W.E.
- Kerensky R.A.,
- Bertolet B.D.,
- Deedwania P.,
- et al.
- Boden W.E.,
- O’Rourke R.A.,
- Dai H.,
- et al.
- Ferry D.R.,
- O’Rourke R.A.,
- Blaustein A.S.,
- et al.
- Bittl J.A.