Considerable Time From the Onset of Plaque Rupture and/or Thrombi Until the Onset of Acute Myocardial Infarction in Humans
Coronary Angiographic Findings Within 1 Week Before the Onset of Infarction
Background—It has been thought that the thrombi and bleeding in plaques that occur after plaque rupture or endothelial damage from vessels with mild stenosis suddenly occlude the lumen and cause acute myocardial infarction (AMI). However, our hypothesis is that thrombi and bleeding may not suddenly occlude the lumen.
Methods and Results—The study group consisted of 20 patients who had coronary angiograms performed within 1 week (3±3 days) before AMI and 20 control patients who had coronary angiograms performed 6 to 18 months (282±49 days) before AMI. The features of infarct-related coronary segments (IRCS) at 3 days before AMI were the presence of a significant stenosis of >50% (95% in incidence and 71±12% diameter stenosis) and Ambrose’s type II eccentric lesions (plus multiple irregularities), an indicator of plaque rupture and/or thrombi (60% [70%]), and the features at 1 year before AMI were mild stenosis of <50% (95% incidence and 30±18% diameter stenosis) with rare Ambrose’s type II eccentric lesions (plus multiple irregularities) (10% [10%]). The same relation was observed in each of the 4 subgroups with Q-wave infarction, non–Q-wave infarction, preceding effort angina within 1 month before AMI, and no preceding effort angina.
Conclusions—The appearance of marked progression and Ambrose’s type II eccentric lesion on coronary angiograms 3 days before AMI suggests the presence of a considerable time from the onset of plaque rupture and/or thrombi until the onset of AMI. These features may be predictors of AMI. The concept provides new insight into the mechanism and prevention of human AMIs.
Plaque rupture or endothelial damage and subsequent occlusive thrombi and bleeding in the plaque are considered to be a pathognomonic mechanism of acute myocardial infarction (AMI).1 2 Preexisting stenosis of the infarct-related coronary segments (IRCSs) has been reported to be mostly mild, and their morphology is smooth vessel wall on the basis of coronary angiograms (CAGs) that were mostly performed 1 year or even 3 months before the onset of AMI.3 4 5 6 7 Recently, based on the findings of CAGs performed at ≈4-month intervals during 1 year, we reported that the process of progression of coronary artery lesions is classified into 2 types: type 1 vessels are characterized by the sudden appearance of marked progression due to large thrombi and bleeding in plaques after plaque rupture or endothelial damage as indicated by Ambrose’s type II eccentric lesions; and type 2 vessels are characterized by the continuous slight progression of stenosis due to plaque growth or small thrombi and bleeding in plaques after plaque rupture or endothelial damage as indicated on smooth vessel walls on CAGs.7 AMI occurred in type 1 vessels, and effort angina pectoris occurred with both type 1 and 2 vessels when the percent diameter stenosis became severe. Thus, it has been generally considered that plaque rupture or endothelial damage suddenly occlude the lumen of the future IRCSs with mild stenosis at the onset of AMI.
However, the residual stenosis was moderate or severe in most patients with AMI who successfully underwent coronary thrombolysis.8 9 10 Moreover, it has been established that about half of patients with AMI had experienced angina within 1 month before AMI development,11 12 13 suggesting the appearance of severe stenosis at the onset of preceding angina. In addition, it is well known that plaque rupture or endothelial damage does not occlude the lumen of angina-related segments in unstable angina. Therefore, we hypothesized that plaque rupture or endothelial damage does not suddenly occlude the lumen of the future IRCS with mild stenosis, although it may further the degree of stenosis to a moderate or severe level. That means that a considerable time may pass from the onset of plaque rupture or endothelial damage to occlusion of the lumen in AMI patients with and without preceding effort angina. To define this hypothesis, CAGs within 1 week before the onset of AMI should be examined. Thus, we designed the present study to retrospectively compare the features of CAGs performed 1 year and within 1 week before the onset of AMI.
A Multicenter Study in Gifu University and Affiliated Hospitals on Cardiac Disease (MUGIC) was organized by 1 university hospital (Gifu University School of Medicine) and 6 affiliated hospitals (Gifu Municipal Hospital, Gifu Prefectural Hospital, Matsunami General Hospital, National Gifu Hospital, and Tosei General Hospital, Gifu Japan; and National Toyohashi-Higashi Hospital, Aichi, Japan). In each of these hospitals, >1000 patients underwent a CAG each year, and 70 to 100 patients with AMI who had undergone an emergency CAG were admitted to each hospital per year.
Patients were selected from computer-stored patient lists at the hospitals and consisted of a total of ≈3000 patients who had experienced an AMI between 1991 and 1997. The criteria for the selection of patients were as follows. (1) Patients had to have had an AMI, which was confirmed on the basis of typical chest pain that continued for >30 minutes, newly developed ischemic ST-T changes or Q waves, and elevation of serum creatine kinase levels to ≥3 times the upper limit of the normal range. (2) Patients had to have undergone an elective but not emergency CAG within 1 week of (group 1) or from 6 to 18 months (group 2) before the AMI episode. (3) IRCSs (objective vessels) were easily identifiable on the CAGs performed at the onset of AMI, and treatments such as CABG, PTCA, or the passage of a guide wire were not carried out in the future IRCS before the onset of AMI. When the quality of CAG was nonsatisfactory or the projection views of 2 CAGs were different, the patients were excluded from the study. Reperfusion with direct PTCA or thrombolysis was performed at the acute phase of AMI in all patients in groups 1 and 2. Twenty patients in group 1 and 20 patients in group 2 met the selection criteria, and the hospitalization records were reviewed for preadmission and admission histories of angina as well as for symptoms at the onset of and during AMI. In all cases, patients had been questioned shortly after admission about anginal chest pain in the month before the occurrence of AMI. The diagnosis of angina was based on the clinical opinion of the admitting physician and guided by World Health Organization criteria.14
Group 1 patients included 10 patients who exhibited preceding effort angina and 10 patients who did not exhibit preceding effort angina in the month before the onset of AMI (Table 1⇓). The patients with preceding effort angina had been treated with β-blockers, calcium channel antagonists, aspirin, and nitrates. For these patients, CAGs were performed within 1 week before AMI for the clinical evaluation of the severity of the effort angina. In all 10 patients, the angina-related vessels that exhibited both the most severe stenosis and >70% diameter stenosis on the CAGs before the onset of AMI became occluded and were considered to be the IRCSs at the onset of AMI. Elective, but not emergency, PTCA treatment of the angina-related vessels was scheduled after the CAGs because the diameter stenosis was >70% in each case. However, an AMI occurred in each patient before elective PTCA was performed.
Five of 10 patients in group 1 who had not exhibited preceding effort angina had undergone PTCA during the 3 months before the onset of AMI for the treatment of severe stenosis of vessels that were not later infarct related, and the patients had no angina-like chest pain after PTCA. They had been treated with warfarin, antiplatelet drugs, or both. The CAGs performed within 1 week before AMI were follow-up studies after PTCA. Restenosis after PTCA was not seen in any patients. Three of the other 5 patients exhibited new-onset rest angina that was probably due to vasospasm; the angina-like chest pain was observed only in the early morning or at midnight during bedrest. The CAGs were performed for the evaluation of chest pain. AMI occurred in these patients despite treatment with vasodilators such as calcium channel antagonists or nitrates and the disappearance of chest pain. The remaining 2 patients had ECG abnormalities, and CAG was performed.
One year before AMI, 15 of 20 patients in group 2 had effort angina and CAG was performed. The 10 patients with effort angina had been treated with PTCA directed against the angina-related segments, which did not become future IRCSs, and with warfarin, antiplatelet drugs, or both. Their effort angina disappeared. The other 5 patients with effort angina had been treated with β-blockers, calcium channel antagonists, and nitrates. Effort angina disappeared in 2 patients and was sometimes observed in the other 3 patients. Three of 5 patients without effort angina had rest angina in the early morning and were diagnosed with vasospastic angina without significant stenosis on CAG 1 year before AMI. They were treated with calcium channel antagonists and nitrates. In the remaining 2 patients without effort angina, CAGs were performed 1 year before AMI because of the ECG abnormalities, and no significant stenosis was shown. In the month before the onset of AMI, preceding effort angina was seen in 7 of 20 patients in group 2, and preceding rest angina in the early morning, suggesting vasospastic angina, occurred in 3 patients. It was considered that the preceding effort angina–related artery was the same as the IRCS in each of the 7 patients, because the percent diameter stenosis of coronary arteries, except for the IRCSs, was <70% on CAG after the onset of AMI.
To exclude the effect of coronary artery spasms, all patients were administered an intracoronary injection of nitroglycerin before undergoing CAG. Right and left coronary CAGs were carried out in 3 and 6 projections. The best projection that represented the stenosis of the lesion was selected for the measurement of percent diameter stenosis. Each of 2 coronary arteriograms performed before and after AMI in a given patient was assessed on 2 side-by-side projectors. The culprit lesion of AMI was identified on CAG after AMI with a reference to the ECG change, ventricular wall motion on left ventriculography and echocardiography, or a defect on myocardial scintigraphy. Coronary artery lumen diameters were measured with a computer-assisted CAG analysis system (CCIP-310; Cathex Company). Significant coronary artery stenosis was defined as a >50% luminal diameter narrowing.
The morphologies of the CAGs were reviewed by 2 observers who were experienced in angiographic interpretation and blinded to the clinical data according to the criteria proposed by Ambrose et al.15 16 Briefly, lesions were classified as concentric lesions, type I eccentric lesions, type II eccentric lesions, or lesions with multiple irregularities. Differences were resolved on consensus between the 2 observers.
Coronary Risk Factors
Risk factors for AMI were identified, including resting blood pressure, smoking status, and the presence of diabetes and hyperlipidemia. These risk factors were identified based on data obtained within 1 month before the onset of AMI. Hyperlipidemia was considered present if the patient was receiving appropriate therapy or if serum total cholesterol was ≥240 mg/dL or LDL cholesterol was ≥160 mg/dL. Hypertension was considered present if the patient was receiving therapy or if systolic blood pressure was ≥140 mm Hg or diastolic blood pressure was ≥90 mm Hg. Diabetes was considered present if the patient was receiving appropriate therapy, if the patient had abnormal measurements in the glucose tolerance test, or if the fasting blood glucose concentration was ≥140 mg/dL.
The unpaired Student’s t test was used to compare continuous data between groups, and χ2 analysis was used to compare categorical data. A value of P<0.05 was considered statistically significant. All data are presented as mean±SD.
Patient Characteristics of Groups 1 and 2
Data are summarized in Table 1⇑. Between groups 1 and 2, there were no significant differences in age, sex, symptoms at 1 month before AMI, location of AMI, frequency of patients with Q- and non–Q wave infarction, death within 1 month after the onset of AMI, location and TIMI grade17 of IRCS at AMI, grade of collateral circulation, lesion of non-IRCS, and incidence of risk factors. However, the time from the onset of AMI to reperfusion was significantly lower in group 1 (2±1 hours) than in group 2 (4±2 hours) because of a higher incidence of hospital onset of AMI in group 1 (14 of 20 patients in group 1 and 0 of 20 patients in group 2).
CAG Findings for Groups 1 and 2
The percent diameter stenosis of the infarct-related segment was 30±18% at 1 year before AMI and 71±12% at 3 days before the onset of AMI and became occlusive after the onset of AMI (TIMI grade 0, 1, 2, and 3, n=23, 4, 12, and 1) (Table 2⇓ and Figures 1 to 3⇓⇓⇓). The difference was significant in each group. The IRCS with a significant stenosis of >50% was observed in 1 of 20 segments (5%) at 1 year before AMI and in 19 of 20 segments (95%) at 3 days before the onset. The difference was significant. The same analysis was performed in each of the 4 subgroups in groups 1 and 2: patients with Q-wave infarction, with non–Q-wave infarction, with preceding effort angina, and without preceding effort angina within 1 month before AMI. As shown in Table 2⇓, the incidence of IRCSs with significant stenosis and the percent diameter stenosis was significantly higher at 3 days before the onset of AMI than at 1 year before AMI in each of the 4 subgroups. There was no evidence of definite regression.
Ambrose’s type II eccentric lesion in future IRCSs was observed in 60% of segments at 3 days before the onset of AMI and in 10% of segments at 1 year before AMI (Table 2⇑). The difference was significant. It was also significantly higher at 3 days before the onset of AMI than at 1 year before AMI in each of the 3 patient subgroups with Q-wave infarction or preceding effort angina and without preceding effort angina. Although the difference did not reach significance in the non–Q-wave infarction group (P=0.170), Ambrose’s type II eccentric lesion plus multiple irregularity was significantly higher at 3 days before the onset of AMI than at 1 year before AMI for the non–Q-wave infarction group (Table 2⇑).
Incidence of AMI for Effort Angina–Related Segments, Segments With Significant Stenosis, and Segments With Ambrose’s Type II Eccentric Lesion
The segments in which the passage of a guide wire or PTCA was performed before AMI were excluded from the study. Data are summarized in Table 3⇓.
In 20 group 1 patients, effort angina within 1 month before AMI was observed in 10 patients (50%). All of 10 effort angina–related segments became occlusive IRCSs 3 days later. Significant diameter stenosis on the CAG 3 days before AMI was seen in 34 segments. Nineteen of 34 segments (56%) became occlusive IRCSs 3 days later. There was no progression of the lesion in the remaining 15 segments on the onset of AMI. Ambrose’s type II eccentric lesion was seen in 14 segments, of which each showed a significant diameter stenosis of >50%, at 3 days before AMI. Twelve of 14 segments (86%) became occlusive IRCSs 3 days later. There was no progression of the lesion in the other 2 segments on the onset of AMI.
In 20 group 2 patients, effort angina 1 year before AMI was observed in 15 patients. Because 10 of 15 angina-related segments were treated with PTCA, these were excluded from the study. The morphology showed Ambrose’s type II eccentric lesion in 1 of the other 5 angina-related segments and in smooth vessel walls in the remaining 4 segments.
In group 2, a significant diameter stenosis of >50% was seen in 15 segments of 20 patients on CAG 1 year before AMI. Only 2 of 15 segments showed Ambrose’s type II eccentric lesions. Only 1 segment with Ambrose’s type II eccentric lesion (7%) became an occlusive IRCS 1 year later. In the other 14 segments with smooth vessel walls, the percent stenosis and morphology were similar during the year. However, Ambrose’s type II eccentric lesions were seen in 3 segments of 20 patients at 1 year before AMI. The 2 segments with and without significant diameter stenosis (78% and 49%) became occlusive IRCSs 1 year later. However, there was no progression in the other segment with significant stenosis (61%).
The present study reveals that a feature of future IRCS was mild stenosis and smooth vessel wall without irregularity 1 year before AMI but moderate or severe stenosis with Ambrose’s type II eccentric lesion 3 days before the onset.
Are Groups 1 and 2 Similar to the Usual AMI?
In the present study, patient age and sex, location of AMI, location and TIMI grade of the IRCSs at AMI, and incidence of risk factors in group 1 were similar to those of group 2 and those of Japanese patients with AMI in general.18 19 20 21 22 However, the time between the attack of AMI and reperfusion was shorter in group 1 than in group 2. This is explained by the higher incidence of the onset of AMI during hospitalization for elective CAG or from the elective CAG to the elective PTCA in group 1. Because CAG was performed 3 days before AMI in group 1, the cause of AMI and the effect of the CAG in patients of group 1 should be discussed, especially in the 6 patients in whom the time from CAG before AMI to AMI was short (2.5 to 8 hours). In the 6 patients, the percent diameter stenosis of the future IRCS was 70% to 94% at CAG before AMI. Five showed Ambrose’s type II eccentric lesions, indicating plaque rupture, thrombi, or both; the sixth patient also showed multiple irregularities. Each of the 6 patients had preceding effort angina that appeared >7 days before AMI, and the angina-related arteries became IRCSs on the onset of AMI, suggesting that plaque rupture had occurred or thrombi of IRCSs had appeared ≥7 days before AMI. Elective, but not emergency, CAG was performed without any complications by experienced cardiologists in each of the 20 group 1 patients. The passage of a guide wire was not performed in any IRCS. No special treatment was performed from the time of CAG to the time of the AMI. In addition, the onset of AMI immediately after elective CAG is rare. These factors suggest that the AMI in group 1 was not related to the preceding CAG.
The incidence of Q-wave infarction was 60% in group 1 and 65% in group 2. The incidence tended to be low in group 1 compared with that of the previous report (71%).23 In general, Q-wave infarction depends on the length of the occlusion time of the IRCS and the inadequacy of collateral circulation. In group 1, the grade of collateral circulation at AMI was similar to that of group 2 and the usual AMI.24 However, the time from the onset of AMI to reperfusion was relatively short in group 1 compared with that in group 2 and Japanese patients with AMI in general. The time was shorter in group 1 patients with non–Q-wave infarction (1.6±1.0 hours) than in group 1 patients with Q-wave infarction (3.1±1.8 hours) (P<0.05). Thus, the low tendency of the incidence of Q-wave infarction in group 1 could be explained by the earlier reperfusion time.25 Also, the incidence of death within 1 month after the onset of AMI tended to be low in group 1 (0 of 20 patients) compared with that in group 2 (2 of 20 patients) and patients with AMI in general (6.3% to 7.4%).26 Earlier reperfusion after the onset of AMI in group 1 may also contribute, because it may improve the prognosis of AMI. These were independent of the onset of AMI.
Most patients in the present study already had coronary heart disease before the onset of AMI. Therefore, the precise relationship between AMI and the future IRCS remains unknown in patients who did not exhibit preceding coronary heart disease. However, the incidence of preceding effort angina within 1 month before the onset of AMI (50% in group 1 and 35% in group 2) was similar to that of patients with AMI in general who had not undergone CAG before the onset of AMI (30% to 60%).11 12 13 27 28 The angina-related vessels were considered the future IRCSs in all patients in groups 1 and 2 (see Methods). The number of diseased vessels with a significant stenosis of >50% except for infarct-related arteries was 0.8±0.8 in group 1 and 0.9±0.8 in group 2, which were similar values as in a previous report.29 These can be explained by the inclusion of patients who previously undergone PTCA for the treatment of stenosed non–future IRCSs in the present study. Thus, it is considered that the group 1 patients and group 2 patients had a similar background on the onset of AMI to the patients with AMI in general.
Progression and Mechanism of the Lesions in Future IRCSs
In the present study, the degree of progression in the future IRCS from 1 year before AMI to 3 days before AMI was marked in percent diameter stenosis. It was also marked in each of the 4 subgroups of Q-wave infarction, non–Q-wave infarction, preceding effort angina, and no preceding effort angina. Ambrose’s type II eccentric lesion was rare in the future IRCSs at 1 year before AMI but was frequently observed in the future IRCSs at 3 days before AMI. The same relation was also seen in each of the 4 subgroups. CAG itself is not a precise method for the detection of plaque rupture or thrombi and in particular underestimates plaque rupture or thrombi with a small size. In addition, the degree of stenosis in the future IRCSs 3 days before AMI was severe in the patients who exhibited effort angina within 1 month before AMI and moderate in patients without it. That is, the presence or absence of preceding effort angina depends on the severity in percent diameter stenosis after marked progression. These suggest that IRCSs in the present study are the type 1 vessels described in the introduction and that the mechanism of marked progression between 1 year before AMI and 3 days before AMI is plaque rupture or thrombi. In addition, the future IRCSs reprogressed into occlusive lesions 3 days later and became IRCSs. Whether the reprogression is due to pathological progression of plaque rupture or changes of blood on thrombogenesis remains unknown.
Duration Between Plaque Rupture and the Onset of AMI
To our knowledge, there is no precise study on the duration from the onset of plaque rupture and/or thrombi to the occlusion of IRCSs at the onset of AMI. The present study revealed that Ambrose’s type II eccentric lesions indicating plaque rupture and/or thrombi and marked progression in percent diameter stenosis were observed 3 days before AMI, suggesting that the plaque rupture and/or thrombi had already occurred. Approximately half of the patients with AMI had preceding effort angina within 1 month before AMI.11 12 13 The mean time of the appearance of chest pain before AMI was ≈7 days. This was confirmed in the present study. In these patients, it was considered that the onset of plaque rupture and/or thrombi was the time at which the chest pain appeared or earlier. However, in patients without preceding effort angina, the mean duration from preceding CAG to the onset of AMI was 4 days in the present study, indicating that the plaque rupture and/or thrombi had occurred earlier. This suggests that the time from the onset of plaque rupture and/or thrombi until the occlusion of IRCS may be considerably longer in most patients both with and without preceding effort angina. Thus, the process of the onset of AMI is classified into 2 steps. Step 1 consists of the onset of plaque rupture and/or thrombi from future IRCSs with mild or moderate stenosis, followed by marked progression to moderate or severe stenosis but not occlusive lesion. Step 2 consists of reprogression from moderate or severe stenosis into occlusive lesion and the appearance of AMI. The concept of step 1, with considerable time, is probably important for the mechanism and prevention of the onset of human AMI.
Is the Onset of AMI Predictable on the Basis of CAG Findings Before AMI?
Until now, it has been considered difficult to predict the onset of AMI even with CAG findings obtained 3 months before the onset of AMI, because most AMIs occur suddenly in vessels with mild or moderate stenosis and smooth vessel walls.3 4 5 6 7 In the present study, the location of the subsequent infarct was also unpredictable on the basis of percent diameter stenosis 1 year before AMI. It also is not clear whether it is predictable from Ambrose’s type II eccentric lesions on CAG 1 year before AMI, because of the small number of segments in the present study.
The present study revealed that a feature of future IRCSs on CAG, 3 days before the onset of AMI, was a significant percent diameter stenosis of >50% and Ambrose’s type II eccentric lesion in patients both with and without preceding effort angina. However, the features of the segments in which significant stenosis were present but for which AMI did not occur for 1 year were smooth vessel wall and rare Ambrose’s type II eccentric lesions. This suggests that the presence of Ambrose’s type II eccentric lesion in the patients with and without preceding effort angina is a predictor for recent AMI, which further suggests that preventive PTCA may be useful for the prevention of the onset of AMI in patients both with and without angina when irregular vessel walls are observed. However, the number of patients in the present study who did not develop AMI despite the appearance of Ambrose’s type II eccentric lesion is unknown. Further investigation is warranted.
In conclusion, most human AMIs do not occur immediately after the onset of plaque rupture and/or thrombi but rather after at least 3 days. Ambrose’s type II eccentric lesion, an indicator of plaque rupture and/or thrombi, probably is a predictor of recent AMI. This concept provides us with fundamental and important information for understanding the mechanism and prevention of the onset of AMI.
- Received March 2, 2000.
- Revision received May 23, 2000.
- Accepted June 8, 2000.
- Copyright © 2000 by American Heart Association
Falk E. Plaque rupture with severe pre-existing stenosis precipitating coronary thrombosis: characteristics of coronary atherosclerotic plaques underlying fatal occlusive thrombi. Br Heart J. 1983;50:127–134.
Davies MJ, Thomas AC. Plaque fissuring—the cause of acute myocardial infarction, sudden ischemic death, and crescendo angina. Br Heart J. 1985;53:363–373.
Little WC, Constantinescu M, Applegate RJ, et al. Can coronary angiography predict the site of a subsequent myocardial infarction in patients with mild-to-moderate coronary artery disease? Circulation. 1988;78:1157–1166.
Dacanay S, Kennedy HL, Uretz E, et al. Morphological and quantitative angiographic analyses of progression of coronary stenoses: a comparison of Q-wave and non–Q-wave myocardial infarction. Circulation. 1994;90:1739–1746.
Yokoya K, Takatsu H, Suzuki T, et al. Process of progression of coronary artery lesions from mild or moderate stenosis to moderate or severe stenosis: a study based on four serial coronary arteriograms per year. Circulation. 1999;100:903–909.
Serruys PW, Arnold AER, Brower RW, et al. Effect of continued rt-PA administration on the residual stenosis after initially successful recanalization in acute myocardial infarction: a quantitative coronary angiography study of a randomized trial. Eur Heart J. 1987;8:1172–1181.
Harper RW, Kennedy G, DeSanctis RW, et al. The incidence and pattern of angina before acute myocardial infarction: a study of 577 cases. Am Heart J. 1979;2:178–183.
Stowers M, Short D. Warning symptoms before major myocardial infarction. Br Heart J.1970;32:833–838.
Expert Committee on Cardiovascular Diseases and Hypertension. Hypertension and coronary heart disease: classification and criteria for epidemiological studies. In: World Health Organization. Technical Report Series. 1959;168:15–16.
Ishikawa K, Nakai S, Takenaka T, et al. Short-acting nifedipine and diltiazem do not reduce the incidence of cardiac events in patients with healed myocardial infarction. Circulation. 1997;95:2368–2373.
Aguirre FV, Younis LT, Chaitman BR, et al, for the TIMI II Investigators. Early and 1-year clinical outcome of patients’ evolving non–Q-wave versus Q-wave myocardial infarction after thrombolysis: results from the TIMI II study. Circulation. 1995;91:2541–2548.
Berger PB, Ellis S, Holmes DR, et al, for the GUSTO-II Investigators. Relationship between delay in performing direct coronary angioplasty and early clinical outcome in patients with acute myocardial infarction: results from the Global Use of Strategies to Open Occluded Arteries in Acute Coronary Syndromes (GUSTO-IIb) trial. Circulation. 1999;100:14–20.
Mounsey P. Prodromal symptoms in myocardial infarction. Br Heart J. 1951;14:215–226.
Solomon HA, Edwards AL, Killip T. Prodromata in acute myocardial infarction. Circulation. 1969;40:463–471.