Author + information
- Received September 1, 2001
- Revision received December 27, 2001
- Accepted February 5, 2002
- Published online May 1, 2002.
- Richard A Kerensky, MD*,* (, )
- Michael Wade, MS†,
- Prakash Deedwania, MD‡,
- William E Boden, MD†,§,
- Carl J Pepine, MD, MACC*,
- Veterans Affairs Non–Q-Wave Infarction Strategies in-Hospital (VANQWISH) Trial Investigators
- ↵*Reprint requests and correspondence:
Dr. Richard A. Kerensky, Division of Cardiovascular Medicine, University of Florida College of Medicine, P.O. Box 100277, Gainesville, Florida 32610-0277, USA.
Objectives We sought to determine the underlying coronary anatomy and characterize the culprit lesion after non–Q-wave myocardial infarction (NQWMI).
Background Although the culprit lesion and infarct-related artery often are easily identified with coronary angiography after Q-wave MI, the culprit lesion after NQWMI has not been well characterized. Small retrospective studies have suggested that the absence of Q-waves on an electrocardiogram is due to incomplete occlusion of the infarct-related artery.
Methods Coronary angiograms from 350 patients randomized to the early invasive strategy in the Veterans Affairs Non–Q-Wave Infarction Strategies in-Hospital (VANQWISH) trial were systematically analyzed in an angiographic core laboratory. A consensus panel identified the culprit lesion and the infarct-related artery using prespecified criteria for complex lesion morphology and acute versus chronic occlusions. Severity of angiographic disease and left ventricular function also were analyzed. Patients with a single identified culprit lesion were compared with those who had multiple apparent culprits and those without an identifiable culprit lesion.
Results A single culprit lesion was identified in only 49% of patients undergoing early angiography after NQWMI. The majority of patients either had no identifiable culprit (37%) or multiple apparent culprit lesions (14%). A single incomplete occlusion of the infarct-related artery was found in only 36% of patients, and an isolated acute occlusion of the infarct-related artery occurred in 13%. Patients without an identifiable culprit lesion had severe coronary disease (obstructive coronary artery disease [CAD] in 84%) but no complex lesion morphology. There was no difference in angiographic severity of disease comparing patients with and without identifiable culprit lesions. Patients with a single incomplete occlusion of the infarct-related artery were more likely to undergo percutaneous transluminal coronary angioplasty than other patients, whereas patients with multiple culprit lesions were more frequently treated with coronary artery bypass grafting.
Conclusions Coronary angiography early after NQWMI frequently identifies severe obstructive CAD, but a single identifiable culprit lesion was identified in <50% of patients. Multiple culprit lesions were seen in 14% of patients. An angiographic culprit lesion could not be identified in more than one-third of patients undergoing coronary angiography as part of an invasive strategy.
The culprit lesion in Q-wave myocardial infarction (MI) is an acute occlusion of a major epicardial vessel easily visualized angiographically (1). Culprit lesions in Q-wave MI are the target of treatment with thrombolytic therapy or primary percutaneous transluminal coronary angioplasty (2–9). These lesions are located at the site of a ruptured atherosclerotic plaque (10,11). In contrast, the culprit lesion in non–Q-wave MI (NQWMI) is not well characterized. The few retrospective angiographic studies of patients after NQWMI have suggested that the infarction is associated with incomplete occlusion of the infarct-related artery (12–14). One of the primary reasons for performing angiography early after MI is to identify the culprit lesion and assess the infarct-related arteries’ suitability for revascularization. The purpose of our study was twofold: 1) to determine how frequently a culprit lesion can be identified as the cause of the MI in patients with NQWMI, and 2) to characterize the coronary anatomy in patients with and without an identifiable culprit lesion.
We separated patients into two groups: those with an identifiable culprit lesion (complex stenosis or acute occlusion) and those without an identifiable culprit lesion. Patients with an identifiable culprit lesion were defined as having incomplete occlusion of the infarct-related artery and complex lesion morphology (acute culprit lesions) or acute occlusion of the infarct-related artery (plaque rupture with occlusive thrombus). Patients without an identifiable culprit lesion were found to have obstructive coronary disease of varying severity but no complex lesion morphology, non-obstructive coronary artery disease (CAD) or angiographically normal coronary arteries.
The Veterans Affairs Non–Q-Wave Infarction Strategies in-Hospital (VANQWISH) trial is the largest prospective, multicenter randomized study yet undertaken of NQWMI. Its primary objective was to compare effects of an early invasive versus non-invasive management strategy on long-term clinical outcome (death or MI) after NQWMI. In this study we sought to identify, angiographically, the cause of the documented NQWMI in patients randomized to the routine early invasive strategy.
The protocol design, baseline characteristics and primary outcome data are published in detail elsewhere (15,16). Data reported here are from patients randomly assigned to the early invasive management strategy who had protocol-mandated coronary angiography and core laboratory readings. Briefly, patients were eligible if they met all the following criteria: clinical presentation of evolving acute MI, MB-CK isoenzyme >1.5 times upper normal limit and no new Q-waves on serial electrocardiograms (ECGs). Reasons for exclusion from the trial were cardiogenic shock, severe heart failure, persistent or recurrent ischemia at rest or ventricular tachyarrhythmias.
All coronary angiograms were done on-site within three to seven days of NQWMI using standard techniques and were recorded on 35-mm cine film. Films were reviewed on-site for clinical decision-making and subsequently sent to an angiographic core lab, where they were read by a panel of experienced angiographers. This report concerns the angiographic core lab data. The core lab readers and technical personnel were blind to all clinical and site data. Angiograms were analyzed qualitatively by a consensus panel. The films were first given an overall quality score of superior, satisfactory or unsatisfactory. Major arteries were defined as the left anterior descending (LAD), the left circumflex (LCx), the right coronary (RCA) and the left main. Left ventricular angiograms also were analyzed qualitatively for left ventricular function. This was done by visual assessment of segmental wall motion and estimation of ejection fraction.
Lesions were localized to coronary segments according to the Coronary Artery Surgery Study definition, and the percent diameter stenosis was estimated from a view demonstrating the most severe narrowing. An attempt was made to identify possible culprit lesions in each patient using angiographic criteria for complex coronary stenosis published elsewhere (17–23). Clinical and ECG data were not provided to the consensus panel. For the purpose of this study, a culprit lesion was defined as a lesion with complex features suggestive of acute plaque rupture. Six criteria were used to describe a culprit lesion: intraluminal filling defect (or thrombus), ulcer with overhanging edges, extraluminal contrast, dissection or intraluminal flap, multiple irregularities or acute occlusion. In lesions with multiple features, the predominant feature was used to classify the lesion by consensus. An occlusion was considered acute if it ended abruptly, with a squared-off or convex upstream termination. An occlusion was considered chronic if it tapered smoothly before the obstruction, or if the exact site of the occlusion was not well visualized. Ultimately, lesions were classified as complex or not complex, and occlusions classified as acute or chronic, based on consensus of the panel using these criteria as guidelines. Collateral filling to the distal vessel beyond the culprit lesion was noted as present or absent. Flow distal to the culprit lesion was visually estimated by the method used in the Thrombolysis In Myocardial Infarction (TIMI) trials and graded on a scale of 0 to 3 (24).
Patients assigned to the invasive strategy who had a superior or satisfactory angiographic film were included in this analysis. Baseline characteristics of these patients were compared with those of the remaining patients in the invasive strategy to ensure there was no selection bias. Categorical baseline factors between these two groups were analyzed using Pearson’s chi-square test, whereas continuous variables were compared by Student ttest. Pearson’s chi-square test also was employed to compare the proportion of patients with angiographically normal coronaries, three-vessel CAD, severe (>70%) proximal stenosis and total (100%) stenosis between those who had no and ≥1 culprit lesion. Patients with no culprit lesion, a single incomplete occlusion, a single abrupt occlusion or multiple culprit lesions were compared with respect to treatment with revascularization (coronary angioplasty or coronary artery bypass grafting [CABG]). These four groups of patients were analyzed using Pearson’s test. All analyses were performed using Statistical Analysis Software version 6.12 (SAS Inc., Cary, North Carolina).
Of the 462 patients randomized to the invasive strategy, 442 had the protocol-mandated coronary angiogram. The remaining 20 did not complete the angiogram, because of patient or physician refusal. The core laboratory received films on 381 patients. Transmittal forms or films were missing for the remaining patients. A total of 84% of the films were judged technically superior in quality, 7.9% were satisfactory and 8.1% were unsatisfactory. Unsatisfactory films were not analyzed. Thus, this report is based on findings from 350 patient films.
Clinical, ECG and left ventricular angiographic findings
Clinical findings for these 350 patients are summarized in Table 1and do not differ significantly from the findings in the 112 patients randomized to this strategy whose films were not evaluated by the angiographic core laboratory, except that core laboratory patients were more likely to be male (98% vs. 93%, p = 0.009) and had a slightly lower incidence of posterior infarction (10% vs. 18%, p = 0.030) compared with those whose films were not evaluated by the core laboratory. This group predominantly consisted of older men (98.3% male) with multiple CAD risk factors.
Patients undergoing invasive strategy had a mean ejection fraction of 41% ± 13% by core laboratory analysis. Regional wall motion abnormalities were identified in 258 of the 350 patients (74%). This consisted of akinesia in 29%, hypokinesia in 40% and dyskinesia in 5%.
Coronary angiographic findings
Angiographic cause of MI
The identified angiographic cause of the MI is shown in Table 2. In all, 270 culprit lesions or acute occlusions were identified in 221 of the 350 patients. Thus, in 221 patients (63%), at least one culprit lesion was identified. A single identifiable culprit lesion was identified in only 49%. In patients with a single culprit lesion, an incomplete occlusion was found in 73% and acute total occlusion was seen in 27%. Multiple culprit lesions or acute occlusions were seen in 14% of patients. A culprit lesion could not be identified in 37% of patients. Most of the patients without an identifiable culprit (84%) had obstructive lesions (≥50% stenosis) identified; only 16% had normal coronaries or non-obstructive CAD (<50% stenosis).
Identifiable culprit lesions
Single culprit with patent infarct-related artery
A single culprit lesion without total occlusion was found in 36% of patients (Table 2). The complex lesion morphologies identified in these culprit lesions are shown in Table 3. The most common morphology was ulcer with overhanging edges (42%), followed by filling defect (22%). TIMI flow distal to the culprit lesion was grade 3 in 65%, grade 2 in 21% and grade 1 in 13%. Culprit lesions were identified in the LAD (33%), LCx (31%) and RCA (36%). These culprit lesions were located proximal or midportion of the LAD, LCx or RCA in 72% and in the distal segment or one of the branches in 28%. Collaterals distal to culprit lesions in patent infarct-related arteries were uncommon, occurring in 16% of patients.
The cause of the MI was found to be acute total occlusion in 20% of the patients in this study (Table 2). In 46 cases, the acute occlusion was isolated and identified as a single culprit causing the MI. In the other 25 instances, the acute occlusion occurred in combination with other complex culprit lesions when multiple culprit lesions were identified (as described later). These occlusions occurred in the LAD in 27%, LCx in 38% and RCA in 35%. Acute occlusion location was in the proximal or midportion of the vessel in 70% and branch or distal location in 30%. Collaterals were common in acutely occluded vessels, occurring in 68%. Overall, 328 total occlusions were identified in 185 patients, but only 71 of these met angiographic criteria for acute occlusion.
Multiple culprit lesions
The cause of the MI appeared to be due to multiple culprit lesions in 14% of patients. Forty-four patients had two acute lesions, and four patients had three angiographic culprit lesions. Five patients had multiple culprit lesions in the same vessel (such as proximal LAD and diagonal branch), whereas 43 patients had culprit lesions in two or more different vascular distributions (RCA and LCx). In these cases, two or more angiographic locations appear to have led to the MI.
No identifiable culprit lesion
Coronary angiography did not delineate the cause of the MI in 37% of the patients. In general, these patients had severe CAD without any angiographic culprit lesion. The vast majority of these patients (84%) had obstructive CAD and 46% had three-vessel CAD. Thus, although stenoses were identified, the cause or “culprit” was not identified by angiography. Twenty (16%) of the 129 patients without an identifiable culprit lesion had either normal angiograms or non-obstructive CAD. Interestingly, 50% of these patients had wall motion abnormalities (nine hypokinesia, one akinesia). Patients with or without an identifiable culprit lesion did not differ in terms of three-vessel CAD (39% culprit vs. 46% no culprit), proximal stenosis location (46% culprit vs. 43% no culprit) or total occlusions (55% culprit vs. 49% no culprit) (p = NS).
Previous coronary bypass surgery
Seventy-six (22%) of patients had undergone previous CABG. A total of 226 bypass grafts were analyzed. Seventy-five percent were patent and 25% were occluded. In patients with previous CABG, a culprit lesion was identified in 58%. The culprit lesion was in a bypass graft in 36 patients and in a native vessel in 10 patients. Two patients had two culprit lesions with one in a native vessel and one in a bypass graft. The most common culprit lesion morphology in bypass grafts was ulcer with overhanging edges (31%), followed by filling defect (26%). Only eight abrupt occlusions of bypass grafts were identified as the culprit lesion.
Revascularization in patients with and without identifiable culprit lesions
Patients with a single identifiable culprit lesion were more likely to be treated with coronary angioplasty. In the 127 patients with a single incomplete occlusion as the culprit lesion, 41% underwent coronary angioplasty, whereas patients with a single acute occlusion underwent coronary angioplasty in 28% of cases. Angioplasty was used less frequently in patients with no identifiable culprit lesion (14%) or multiple culprit lesions (15%) (p < 0.001, single culprit vs. no or multiple culprits; Fig. 1).
Patients with multiple culprit lesions were more likely to be treated with CABG than patients with a single or no identifiable culprit lesion. Forty-eight patients had multiple culprit lesions identified, and 56% of these patients underwent CABG. In contrast, only 17% of patients with no identifiable culprit lesion had CABG. Patients with single culprit lesions had intermediate rates of CABG (Fig. 1). Thus, patients with no identifiable culprit lesion had the lowest rate of revascularization with either coronary angioplasty or CABG (31%), whereas patients with multiple culprit lesions were revascularized most often (71%).
This is the first prospective, large-scale detailed coronary angiographic analysis of contemporary patients with NQWMI to be published. In this study, angiograms were analyzed at a central core laboratory, and the focus was to identify a culprit lesion and assess the severity of disease. Our findings indicated that NQWMI is not caused entirely by incomplete occlusion of the infarct-related artery, as has been previously suggested. A single complex stenosis without occlusion was found in only 36% of patients. A single acute total occlusion was also uncommon, found in only 13%. Total occlusions were common in this study, but most appeared angiographically to be chronic. Because 14% of patients had multiple apparent culprit lesions, a single identifiable culprit lesion was found in less than one half of the patients. Many patients, therefore, did not have an identifiable culprit lesion in this study using predefined criteria for acute or complex lesions. The majority of patients without identifiable culprit lesions had severe coronary stenoses, but without angiographic morphologies indicating acute plaque rupture.
Comparison to previous studies
These findings contrast with those of previous studies of NQWMI that presumed a culprit lesion almost always can be identified (12,20,25). Previous angiographic studies of patients with NQWMI are case series of patients undergoing angiography at various time intervals after infarction (12). Most studies have been retrospective and have had relatively small numbers of patients. The indication for and timing of the angiogram vary according to practice patterns. In addition, most previous studies took place before currently used therapies for acute ischemic syndromes, including aspirin, heparin, beta-blockers and calcium channel blockers. In 1986, DeWood retrospectively analyzed the coronary angiograms of 341 patients studied within seven days of NQWMI (12). The patients were grouped on the basis of the time between infarction and angiography. The authors found that the incidence of total occlusion increased over time, from 26% in the <24-h group to 42% in the >72-h group. Complex morphology consistent with a culprit lesion was not evaluated. Data such as these led to the notion that NQWMI is due to incomplete occlusions at risk for progression to total occlusion.
Comparison with Q-wave infarction and unstable angina
Patients with Q-wave infarction often have an easily identifiable single culprit lesion in the artery corresponding to ST segment elevation and/or Q waves (1,2,26). Some of the difference in culprit lesion detection comparing our study to studies of patients with Q-wave MI may be due to earlier angiography in the patients with Q-wave MI. However, even when angiography is delayed after Q-wave MI (after thrombolysis, for example) the culprit is usually easily identifiable as a complex plaque in one of the three coronary territories (6,8). Earlier angiography may have increased our rate of culprit lesion detection, but our data contrast with those of previous studies of Q-wave MI where a clear-cut single culprit is almost always identified (2,3,5,6,8).
In patients with unstable angina it is often assumed that a single culprit lesion can be identified with angiography. Many studies do not include core laboratory analysis or report the number of patients without a definite culprit lesion. Strict criteria for culprit lesion morphology often were not prespecified in these previous studies. Angiographic morphologies consistent with acute plaque rupture have been studied extensively. Ambrose and colleagues identified an Ambrose type II lesion (ulcer with overhanging edges) in 70% of patients with unstable angina undergoing angiography (19). This morphology was seen in 35% of patients with a culprit lesion and in only 23% of the 350 patients undergoing angiography in our study. Culprit lesion morphology in NQWMI was assessed in a central core laboratory in TIMI IIIA, which evaluated the culprit lesion in 306 patients with unstable angina or NQWMI within 12 h of enrollment (20). The majority of patients in this study had unstable angina. A NQWMI was diagnosed in 32% of patients. A culprit lesion was identified in all of the 306 patients entered in TIMI IIIA, but complex lesion morphology was not required for a lesion to be classified as a culprit. Complex lesion morphology was identified in only 215 (70%) of the patients and two culprit lesions were seen in 32% of patients in TIMI IIIA. We identified complex lesion morphology (excluding acute occlusions) in 50% of patients and two or more culprit lesions in 14%, including only patients with NQWMI. Core angiographic analyses from the PRISM-PLUS trial are available for 1,491 patients with unstable angina (56%) and NQWMI (44%) (27). All the patients in this trial had an identifiable culprit lesion, despite the fact that 10% had nonsignificant disease. A stenosis was classified as the culprit even if no complex morphologic features were present. A possible or definite thrombus was seen in 45% of patients in that trial, but other lesion morphologies were not reported. In a recently reported FRISC II substudy, a culprit lesion was identified by local reading in 56% of patients with an acute coronary syndrome, similar to our findings (28).
No or multiple culprit lesions
We do not know why many patients in this study did not have an identifiable culprit lesion. Certainly some of the stenoses identified without complex features were actually culprits, but with so many obstructive lesions present we cannot assume that all or even most of these stenoses were culprit lesions. We speculate that many patients had underlying pathology that could not be detected with a conventional coronary angiogram. Small-vessel occlusion in vessels beyond the resolution of current angiographic techniques is one possibility. Also, some large epicardial vessels may have intraluminal or vessel wall pathology as a stimulus for platelet and thrombus generation without definite features of plaque rupture angiographically. Some culprit lesions may have resolved or “smoothed out” with antithrombotic therapy before angiography. Some patients with NQWMI may not have classic epicardial vessel plaque rupture and thrombosis but, instead, have severe ischemic injury based on increased demand in a setting of limited supply due to severe chronic CAD.
The patients with multiple angiographic culprit lesions are an interesting subgroup (29). These patients may have diffuse activation of multiple plaques with multiple plaque ruptures, similar to what is seen in patients with vasculitis. Alternatively, an initial plaque rupture may initiate a cascade of other triggers, leading to a second or third plaque rupture sequentially. Some complex lesions seen in these patients may have been the result of previous “old” plaque ruptures that healed with a complex morphology. Clearly, patients with NQWMI represent a heterogeneous patient population clinically, and the angiographic findings in this study suggest heterogeneity in the angiographic findings and probably the underlying pathology.
Our study included only patients with NQWMI. We did not assume that a culprit lesion always would be identified. Our data indicate that an early coronary angiogram does not always delineate the cause of the MI in patients with NQWMI. This contrasts sharply with findings from previous studies of patients with Q-wave MI and studies combining patients with unstable angina and NQWMI. Our findings are important because most aggressive strategies are predicated on the assumption that a culprit lesion can be identified and then targeted for revascularization. Aggressive strategies using coronary angioplasty for acute ischemic syndromes report success rates based on revascularization of the culprit lesion or infarct-related artery. In this study of NQWMI, coronary angioplasty was performed in 38% of patients when a single culprit lesion was found and only 14% of patients when no or multiple culprit lesions were seen. These data suggest that although the coronary angiogram is still the gold standard for the identification of coronary artery obstruction, angiographic findings early after NQWMI often will fail to identify the cause of the infarction. The results from the recently reported TACTICS trial are consistent with our findings (30). This trial compared an early invasive strategy to conservative therapy in patients treated with aspirin, heparin and tirofiban. In the invasive arm of this trial, only 41% of the patients were treated with percutaneous intervention. Although no core angiographic laboratory data are available, this low rate of percutaneous revascularization suggests that a single culprit lesion was not frequently identified.
The present study has several advantages over previous reports. A common protocol was used prospectively to select and treat patients using contemporary therapy (e.g., aspirin, heparin, diltiazem, etc.) Angiograms were mandated by protocol and done three to seven days after infarction. A consensus panel of three angiographers who were blind to clinical data analyzed the angiograms. In contrast to previous reports, not all total occlusions were assumed to be culprits and angiographic features of acute occlusion were required for culprit lesions. Morphologies accepted as markers of acute plaque rupture and threatened thrombotic occlusion were prospectively decided upon and included as potential culprit lesions. Thus, this report involves the largest group of patients prospectively undergoing a protocol-mandated coronary angiogram in the era of modern therapy for acute ischemic syndromes.
Several limitations of this analysis deserve attention. First, a general qualitative assessment was performed. This, however, was done using a consensus among three experienced angiographers and is in keeping with other reports. Quantitative computer-assisted angiographic analysis was not performed and would have been very difficult, considering the severity and complexity of disease. Second, ECG data that may help determine the culprit lesion location were not available to core laboratory angiographers. Thus, it is possible our assessment of culprit lesions underestimates their true frequency in these patients. On the other hand, clinicians who are biased by electrocardiographic and clinical information may “overinterpret” the angiogram and assume a chronic stenosis is an acute lesion. Third, the time between presentation and coronary angiography ranged from three to seven days. Accordingly, it is assumed that these angiographic findings were present at the onset of the NQWMI. Fourth, this was a predominantly male population, so the findings may not apply to women with NQWMI. Finally, patients had to be clinically stable for three days before entering this trial; therefore, patients who died or developed reinfarction, recurrent ischemia, heart failure, ventricular tachyarrhythmia or shock during this early time period were not included.
Coronary angiography early after NQWMI frequently identifies severe obstructive CAD, but a single identifiable culprit lesion was identified in <50% of patients. Multiple culprit lesions were seen in 14% of patients and an angiographic culprit lesion could not be identified in more than one third of patients undergoing coronary angiography as part of an invasive strategy. These data should be considered in future studies of revascularization of the infarct-related artery in patients with NQWMI.
Angiographic core laboratory consensus panel
Richard A. Kerensky, MD, Mark Mines, MD, Jay Schlaifer, MD.
The authors thank Melanie Fridl Ross, M.S.J., E.L.S., for the editorial assistance she provided in the preparation of this manuscript.
☆ Supported by a research grant from the Department of Veterans Affairs Cooperative Studies Program and by an unrestricted research grant from Hoechst Marion Roussel.
- coronary artery bypass grafting
- coronary artery disease
- left anterior descending artery
- left circumflex artery
- myocardial infarction
- non–Q-wave myocardial infarction
- right coronary artery
- Thrombolysis In Myocardial Infarction
- Veterans Affairs Non–Q-Wave Infarction Strategies in-Hospital study
- Received September 1, 2001.
- Revision received December 27, 2001.
- Accepted February 5, 2002.
- American College of Cardiology Foundation
- Little W.C,
- Constantinescu M,
- Applegate R.J,
- et al.
- Chesebro J.H,
- Knatterud G,
- Roberts R,
- et al.
- Brown B.G,
- Gallery C.A,
- Badger R.S,
- et al.
- PACT Investigators,
- Ross A.M,
- Coyne K.S,
- Reiner J.S,
- et al.
- Thrombolysis and Angioplasty in Myocardial Infarction Study Group,
- Topol E.J,
- Califf R.M,
- George B.S,
- et al.
- Neuhaus K.-L,
- Feuerer W,
- Jeep-Tebbe S,
- Niederer W,
- Vogt A,
- Tebbe U
- Levin D.C,
- Fallon J.T
- Davies M.J,
- Thomas A.C
- Ferry D.R,
- O’Rourke R.A,
- Blaustein A.S,
- et al.
- Boden W.E,
- O’Rourke R.A,
- Crawford M.H,
- et al.
- Ambrose J.A,
- Winters S.L,
- Arora R.R,
- et al.
- Ambrose J.A,
- Winters S.L,
- Arora R.R,
- et al.
- Ambrose J.A,
- Winters S.L,
- Stern A,
- et al.
- The TIMI IIIA Investigators
- Rehr R,
- Disciascio G,
- Vetrovec G,
- Cowley M
- Freeman M.R,
- Williams A.E,
- Chisholm R.J,
- Armstrong P.W
- Chen L,
- Chester M.R,
- Crook R,
- Kaski J.C
- The TIMI IIIB Investigators
- PRISM-PLUS Investigators,
- Zhao X.Q,
- Theroux P,
- Snapinn S.M,
- Sax F.L
- FRISC II Investigators,
- Lindahl B,
- Diderholm E,
- Lagervist B,
- Venge P,
- Wallentin L
- TACTICS (Treat Angina with Aggrastat and Determine Cost of Therapy with an Invasive or Conservative Strategy) Thrombolysis in Myocardial Infarction 18 Investigators,
- Cannon C.P,
- Weintraub W.S,
- Demopoulos L.A,
- et al.