Diagnostic and Prognostic Value of Holter-Detected ST-Segment Deviation in Unselected Patients With Chest Pain Referred for Coronary Angiography*

Diagnostic and Prognostic Value of Holter-Detected ST-Segment Deviation in Unselected Patients With Chest Pain Referred for Coronary Angiography* A Long-term Follow-up Analysis Chandra K. Nair, MD, FCCP; Ijaz A. Khan, MD; Dennis J. Esterbrooks, MD; Kay L. Ryschon, MS; Daniel E. Hilleman, PharmD Objective: To evaluate the diagnostic and prognostic significance of ST-segment deviation detected by ambulatory Holter monitoring in unselected chest pain patients referred for coronary angiography. Methods: Two hundred seventy-seven patients (71% were men) who underwent coronary angiography for evaluation of chest pain were studied with 24-h ambulatory Holter monitoring within 72 h of angiography. A lumen diameter reduction of > 50% was considered coronary artery disease. The ST-segment deviation was defined as > 1-mm deviation from the baseline lasting > 1 min separated by a minimum of 1 min. The patients were followed up for 65 21 months (mean SD) for occurrences of death, myocardial infarction, hospitalization for unstable angina, and need for revascularization. Results: Of the 277 patients, 223 (80%) had coronary artery disease. The prevalence of coronary artery disease was not significantly different in patients with (43 of 48 patients; 90%) and without (180 of 229 patients; 79%) Holter-detected ST-segment deviation. The diagnostic accuracy of Holter-detected ST-segment deviation in predicting the presence of coronary artery disease was poor (33%), with a sensitivity of 19% and a specificity of 91%. The presence of Holter-detected ST-segment deviation was not predictive of future cardiac events or death. Conclusion: The ST-segment changes detected on ambulatory Holter monitoring are of limited value in identifying coronary artery disease and predicting the future adverse cardiac events or death in unselected patients with chest pain. (CHEST 2001; 120:834 839) Key words: ambulatory ECG monitoring; coronary artery disease; diagnostic accuracy; ECG; Holter monitoring; ischemic heart disease; sensitivity; silent ischemia; specificity; ST segment Abbreviation: NS not significant The ST-segment deviation detected by ambulatory ECG monitoring correlates well with the other objective measures used to define myocardial ischemia in patients with known coronary artery disease. 1 11 The presence of ST-segment deviation during ambulatory ECG (Holter) monitoring in selected patients with stable angina pectoris, unstable angina, myocardial infarction, and aborted sudden cardiac death is also predictive of future cardiac events. 9 16 However, limited data are available regarding the diagnostic and prognostic value of the *From the Division of Cardiology, Department of Medicine, Creighton University School of Medicine, Omaha, NE. Manuscript received December 4, 2000; revision accepted March 21, 2001. Correspondence to: Chandra K. Nair, MD, FCCP, Creighton University Cardiac Center, 3006 Webster St, Omaha, NE 68131 Holter-detected ST-segment deviation in the unselected patients with chest pain referred for coronary angiography. Similarly, the diagnostic significance of the ST-segment deviation detected by ambulatory ECG monitoring is unknown in patients with baseline ST-segment changes known to affect ST-T wave interpretation, whereas exercise-induced ST-segment deviation is known to have a lower diagnostic yield in these patients. 16 In this study, the diagnostic and prognostic significance of Holter-detected ST-segment deviation in unselected patients with chest pain referred for coronary angiography was determined, and was correlated with the extent and severity of coronary artery disease. The diagnostic and prognostic significance of Holter-detected ST-segment deviation was also compared between patients with baseline ST- 834 Clinical Investigations

segment changes known to affect ST-T wave interpretation and those without such changes. Patient Population Materials and Methods Two hundred seventy-seven consecutive patients with chest pain referred for coronary angiography at the Creighton University Medical Center were enrolled in this prospective study. All patients were characterized according to age, gender, history of myocardial infarction, previous revascularization procedures, standard risk factor assessment, concomitant disease status, drug therapy, and angina history. Patients were instructed to continue taking all of their medications, including digoxin and antianginal medications, during ambulatory ECG monitoring. Study Protocol The study was approved by the Institutional Review Board for Human Subjects Research of the Creighton University, and written informed consent was obtained from all participants. All study participants underwent continuous 24-h ambulatory ECG monitoring within 72 h after the coronary angiography. Coronary Angiography: Coronary angiography was performed using the Judkin s approach. Multiple coronary angiograms were made in a variety of projections to ensure the lesion visibility and accuracy. Coronary stenosis was assessed visually and interpreted by two observers unaware of the results of the ambulatory ECG recordings. Significant coronary stenosis was defined as a decrease in lumen diameter of 50% in one or more of the major epicardial coronary arteries or their primary branches. Ambulatory Monitoring: Continuous 24-h ambulatory ECG recordings were made on new ferrous oxide tapes using timecoded, reel-to-reel recorders (model 447; Delmar Avionics; Irvine, CA). The frequency range from these amplitude-modified recorders was 0.05 to 100 Hz. Calibration was accomplished with 1-mV and 0.1-mV square waves using a multiphase calibrator simulating the ECG QRS-complex. A two-lead system (modified lead II and V 5 ) was used with meticulous skin preparation. Electrode sites were verified for lead placement and signal amplitude using a transtelephonic modulator transmitter. Subjects kept a detailed diary of the activities performed and symptoms experienced during the 24-h monitoring period. Ambulatory ECG recordings were reviewed using a Cardio-Data Prodigy Scanner (Mortara Instrument; Milwaukee, WI) at 120 times real time. Significant ST-segment depression or elevation was defined as 1 mm ST-segment deviation occurring 80 ms after the J point, lasting for 60 s. The episodes of ST-segment deviation agreed on by two reviewers were considered as significant. Separation of one episode from another required that the ECG return to baseline for a minimum of 1 min. The total number of the ST-segment deviations and the total duration of the episodes were determined for each patient. The heart rates at the onset and at the time of maximal ST-segment deviation were recorded for each individual episode. The physicians performing ECG analyses were blinded to the coronary angiographic findings and clinical characteristics of the patients. Follow-up: Patients were prospectively followed up for a mean SD of 65 21 months from the time of initial monitoring for occurrences of death, nonfatal myocardial infarction, hospitalization for unstable angina, and need for revascularization with coronary artery bypass graft surgery, percutaneous transluminal coronary angioplasty, or any other transcatheter revascularization procedure. Hospital records, outpatient clinic records, and patient, family or primary physician s interviews were used for confirmation of the events. Deaths were classified as cardiac or noncardiac. Myocardial infarction was diagnosed if two of the following three prospectively defined criteria were met: (1) ischemic type chest pain with duration 20 min; (2) a diagnostic ECG, which consisted of ST-segment elevation of 1 mm in two or more anatomically contiguous leads or development of new left bundle branch block; (3) serum creatine kinase-mb level of 5 ng/ml with a creatine kinase-mb relative index of 2.5% in two consecutive serum samples. Unstable angina was diagnosed if ischemic type chest pain lasted for 20 min in absence of the ECG or enzyme changes consistent with myocardial infarction. Statistical Analysis Data were presented as mean SD where appropriate. A statistical probability of 0.05 was considered significant. The diagnostic accuracy of the ST-segment deviation was examined for all patients and for subgroups of patients with and without baseline ST-segment changes. The prognostic value of the STsegment deviation was examined for all patients, Comparisons of the groups with and without ST-segment deviation were performed using the Student s t test or Pearson s 2 where appropriate. All p values were corrected for multiple comparisons where appropriate. The diagnostic value of the presence of the ST-segment deviation for the detection of coronary artery disease was evaluated by calculation of sensitivity, specificity, positive and negative predictive values, and diagnostic accuracy. The following calculations were used: sensitivity (%) 100 (true-positives)/ (true-positives false-negatives); specificity (%) 100 (truenegative)/(true-negatives false-positives); positive predictive value (%) 100 (true-positives)/(true-positives false-positives); and negative predictive value (%) 100 (true-negatives)/(true-negatives false-negatives); and diagnostic accuracy (%) 100 (true-positives true-negatives/total number of tests). All the statistical analyses were performed using computer software (SPSS version 7.0; SPSS; Chicago, IL). Results Clinical Characteristics The study group consisted of 277 patients (196 men), with a mean age of 63 10 years (range, 32 to 84 years). Clinical characteristics of the patients with and without ST-segment deviation during Holter monitoring are summarized in Table 1. The distribution of baseline characteristics was not different between the two study groups. Seventy-one patients had preexisting conditions affecting the ST segment; 44 patients were receiving digoxin, 1 patient had complete left bundle-branch block, 8 patients had left ventricular hypertrophy, 6 patients had permanent pacemakers implanted, and 12 patients had one or more of the above. Angiographic Findings Two hundred twenty-three patients (80%) had at least one vessel with 50% obstruction, and 54 CHEST / 120 / 3/ SEPTEMBER, 2001 835

Table 1 Baseline Clinical Characteristics of Patients With and Without Holter-Detected ST-Segment Deviation* Characteristics With ST-Segment Deviation (n 48) Without ST- Segment Deviation (n 229) Age, yr 65.6 11.4 62.9 10.2 Men 33 (69) 163 (71) Unstable angina 6 (13) 37 (16) Prior myocardial infarction 6 (13) 28 (12) Hypertension 22 (46) 117 (51) Congestive heart failure 7 (15) 30 (13) Diabetes mellitus 12 (25) 32 (14) Hypercholesterolemia 19 (40) 84 (37) Smoking Active smoking at the time 6 (13) 45 (20) of enrollment Prior or active smoking at the time of enrollment 27 (56) 151 (66) *Data are presented as No. of patients (%) or mean SD; p NS for all characteristics. patients (20%) had no significant coronary obstruction. Single-vessel, double-vessel, and triple-vessel disease was present in 92 patients (33%), 59 patients (21%), and 72 patients (26%), respectively. The prevalence of coronary artery disease in patients with ST-segment deviation was not significantly different than that in patients without ST-segment deviation (43 of 48 patients [90%] vs 180 of 229 patients [79%]; p not significant [NS]). Similarly, there was no significant difference in the single-vessel, doublevessel, and triple-vessel disease between groups (Table 2). Ambulatory Holter ECG Findings Forty-eight patients (17%) had one or more episodes of ST-segment deviation with a total of 120 episodes. Thirty-four patients had 90 episodes of ST-segment depression, and 17 patients had 30 episodes of ST segment elevation. The average duration of episodes of ST depression was 42 60 min Table 2 Coronary Angiographic Findings in Patients With and Without Holter-Detected ST-Segment Deviation* Angiographic Findings With ST-Segment Deviation (n 48) Without ST-Segment Deviation (n 229) Coronary artery disease 43 (90) 180 (79) Single-vessel disease 11 (23) 73 (32) Double-vessel disease 15 (31) 44 (19) Triple-vessel disease 15 (31) 57 (25) Left main disease 2 (4) 6 (3) *Data are presented as No. of patients (%); p NS for all angiographic findings. Table 3 Diagnostic Accuracy of Holter-Detected ST- Segment Changes for Predicting Coronary Artery Disease in all Patients* Variables CAD 1 VD (range, 1 to 272 min). The magnitude of ST-segment depression ranged from 1.0 to 4.2 mm (mean, 2.0 0.8 mm), and it was 1 mm in 21 patients and 2 mm in 13 patients. The average duration of ST-segment elevation episodes was 17 21 min (range, 2.5 to 75.5 min). Magnitude of ST-segment elevation ranged from 1.0 to 6.0 mm (mean, 2.1 1.5 mm). Diagnostic Accuracy of Holter-Detected ST-Segment Deviation The diagnostic accuracy of the ST-segment deviation detected by the ambulatory ECG monitoring in predicting the presence of coronary artery disease for all patients is given in Table 3, and in patients with or without baseline ST-segment abnormalities in Table 4. Of the 223 patients with coronary artery disease, 43 patients (24%) had ST-segment changes. Of the 54 patients without coronary artery disease, 5 patients (10%) had ST-segment changes. Correlation of the Holter-detected ST-segment deviation and the severity and location of obstructed coronary arteries is shown in Table 5. Prognostic Value of Holter-Detected ST-Segment Deviation 2 VD 3 VD Any ST-segment changes Sensitivity 19 13 25 21 Specificity 91 91 91 91 Positive predictive value 90 69 75 75 Negative predictive value 21 40 53 46 Diagnostic accuracy 33 44 57 51 ST-segment elevation Sensitivity 7 6 12 4 Specificity 98 98 98 98 Positive predictive value 94 83 88 75 Negative predictive value 20 40 51 43 Diagnostic accuracy 25 42 53 44 ST-segment depression Sensitivity 14 10 15 17 Specificity 93 93 93 93 Positive predictive value 88 67 69 75 Negative predictive value 21 40 50 46 Diagnostic accuracy 29 42 52 49 *Data are presented as %; p NS for all parameters. 1 VD onevessel disease; 2 VD two-vessel disease; 3 VD three-vessel disease; CAD coronary artery disease. Presence of any ST-segment deviation (depression or elevation) was not significantly correlated with the future cardiac events or death (Table 6). During a 836 Clinical Investigations

Table 4 Diagnostic Accuracy of Holter-Detected ST-Segment Changes for Predicting Coronary Disease in Patients With vs Without Baseline ST-Segment Changes* CAD 1VD 2VD 3VD With Without With Without With Without With Without Any ST-segment change Sensitivity 31 15 18 12 41 19 32 16 Specificity 92 90 92 90 92 90 92 90 Positive predictive value 95 86 75 67 88 67 88 67 Negative predictive value 23 21 46 39 55 52 44 47 Diagnostic accuracy 30 42 42 50 54 63 49 54 ST-segment elevation Sensitivity 9 7 6 6 18 10 5 4 Specificity 92 100 92 100 92 100 92 100 Positive predictive value 83 100 50 100 75 100 50 100 Negative predictive value 19 21 43 39 46 52 36 46 Diagnostic accuracy 25 24 42 43 54 50 47 37 ST-segment depression Sensitivity 22 10 12 9 24 12 27 12 Specificity 100 90 100 90 100 90 100 90 Positive predictive value 100 81 100 60 100 56 100 60 Negative predictive value 22 20 46 38 50 50 45 46 Diagnostic accuracy 26 37 40 50 51 57 47 54 *Data are presented as %; p NS for all parameters. See Table 3 for expansion of abbreviations. mean follow-up of 65 21 months, the incidence of cardiac events (cardiac death, nonfatal myocardial infarctions, hospitalization for unstable angina, and need for revascularization) was similar in patients with and without ST-segment deviation, and was not a function of presence or absence of baseline STsegment abnormalities. Discussion Experience with ambulatory ECG monitoring for myocardial ischemia has been associated with a highly variable sensitivity and specificity for the detection of coronary artery disease. 1 16 The STsegment changes consistent with ischemia have been observed during ambulatory ECG monitoring in 18 to 59% of the patients with known coronary artery disease, and have generally correlated well with the other objective tests used to detect myocardial ischemia. 13 21 The highest prevalence of ambulatory ECG monitoring detected ST-segment changes have been reported in patients with multivessel disease, unstable angina, or a recent myocardial infarction. 22 29 However, the diagnostic accuracy of the exercise stress testing in the patients with coronary artery disease is significantly greater than that observed with the ambulatory ECG monitoring. 17 The sensitivity of the ambulatory ECG monitoring-detected ST-segment changes for detection of myocardial ischemia is low even in patients with triple-vessel Table 5 Correlation of Holter-Detected ST-Segment Changes With Severity of Disease and Coronary Artery Involved* Variables Any ST-Segment Change ST-Segment Elevation ST-Segment Depression Severity of coronary stenosis 50% and 75% 2/30 (7) 1/30 (3) 1/30 (3) 75% and 90% 6/31 (19) 4/31 (13) 2/31 (7) 90% 36/161 (22) 11/161 (7) 28/161 (17) Location of coronary artery stenosis LM 5/24 (21) 2/24 (8) 3/24 (13) LAD/DIAG 17/72 (24) 8/72 (11) 10/72 (14) LCX/OM 15/55 (27) 5/55 (9) 11/55 (20) RCA/PDA 6/34 (18) 2/34 (6) 4/34 (12) *Data are presented as No. of patients/total (%); p NS for all parameters. LM left main; LAD/DIAG left anterior descending/diagonal; LCX/OM left circumflex/obtuse marginal; RCA/PDA right coronary artery/posterior descending artery. CHEST / 120 / 3/ SEPTEMBER, 2001 837

Table 6 Correlation of Presence or Absence of Holter-Detected ST-Segment Changes With Adverse Cardiovascular Outcomes* Variables Cardiac Death Myocardial Infarction CABG PTCA Unstable Angina Combined Outcomes Any ST-segment deviation 7/48 (15) 6/48 (13) 9/48 (19) 8/48 (17) 6/48 (13) 23/48 (48) No ST-segment deviation 23/229 (10) 28/229 (12) 29/229 (13) 36/229 (16) 37/229 (16) 101/229 (44) ST-segment elevation 1 mm 2/17 (12) 4/17 (24) 3/17 (18) 4/17 (24) 3/17 (18) 10/17 (59) ST-segment depression 1 mm 5/34 (15) 3/34 (9) 6/34 (18) 5/34 (15) 3/34 (9) 15/34 (44) *Data are presented as No. of patients/total (%); p NS for all correlations; CABG coronary artery bypass grafting; PTCA percutaneous transluminal coronary angioplasty. or left main coronary artery disease. Presumably, this is because these patients are relatively inactive to avoid angina, and so their heart rates (and double products) are relatively low. This is undoubtedly a major limitation of ambulatory ECG monitoring as opposed to exercise stress testing in detecting myocardial ischemia. The prevalence of the ST-segment changes detected during the ambulatory ECG monitoring in our patients with coronary artery disease was 24%. Although the specificities were high, the sensitivities and diagnostic accuracy of the ST-segment change during ambulatory ECG monitoring in our population were poor. This is generally consistent with other published reports. 12 21 In our study, however, the presence of multivessel disease, unstable angina, or myocardial infarction was not associated with a higher prevalence of ambulatory ECG monitoring detected ST-segment changes, and ambulatory ECG monitoring also failed to predict the severity and location of the coronary artery stenosis, which is in discrepancy with some other published reports. 12 21 These discrepancies between the results of our study and with other published reports may be due to the fact that our patient population consisted of unselected patients of chest pain. Despite the relatively poor and erratic diagnostic value of the ambulatory ECG monitoring, the presence of the Holter-detected ST-segment changes in certain patient populations is highly predictive of future adverse cardiac events. In patients with stable angina, ST-segment changes detected by ambulatory ECG monitoring is an independent predictor of the adverse clinical events including death, myocardial infarction, and revascularization. 12 21 In addition, the ambulatory ECG monitoring provides prognostic information additional to that derived from the established parameters obtained during exercise testing. 13,18,29 Ischemia detected by the ambulatory ECG monitoring in patients with unstable angina or myocardial infarction has also been shown to be predictive of the future adverse clinical outcomes. 25 29 However, the presence of ST-segment changes in our population was not predictive of future adverse clinical outcomes. We believe this occurred as our population consisted of unselected patients with chest pain. Conclusion The ST-segment changes detected on ambulatory ECG monitoring in unselected patients with chest pain are of limited value in identifying the coronary artery disease and predicting the future adverse cardiac events including death, myocardial infarction, need for hospitalization for unstable angina, and revascularization. References 1 Campbell S, Barry J, Rocco MB. Features of the exercise test that reflect the activity of ischemic heart disease out of hospital. Circulation 1986; 74:72 80 2 Mulcahy D, Keegan J, Crean P, et al. Silent myocardial ischemia in chronic stable angina: a study of its frequency and characteristics in 150 patients. Br Heart J 1988; 60:417 423 3 Panza JA, Quyyumi AA, Diodati JG, et al. Prediction of the frequency and duration of ambulatory myocardial ischemia in patients with stable coronary artery disease by determination of the ischemia threshold from exercise testing: importance of the exercise protocol. J Am Coll Cardiol 1991; 17:657 663 4 Valle GA, Lemberg L. Silent ischemia: a clinical update. Chest 1990; 97:186 191 5 Deanfield J, Shea M, Ribiero P. Transient ST-segment depression as a marker of myocardial ischemia during daily life. Am J Cardiol 1984; 54:1195 1200 6 Hirzel H, Leutwyler R, Krayenbuehl H. Silent myocardial ischemia: hemodynamic changes during dynamic exercise in patients with proven coronary artery disease despite absence of angina pectoris. J Am Coll Cardiol 1985; 6:275 284 7 Levy R, Shapiro LM, Wright C. The hemodynamic significance of asymptomatic ST-segment depression assessed by ambulatory pulmonary artery pressure monitoring. Br Heart J 1986; 56:526 530 8 Selwyn A, Allen R, L Abbotte L. Relationship between regional myocardial uptake of rubidium-82 and perfusion: absolute reduction of cation uptake in ischemia. Am J Cardiol 1982; 50:112 121 9 Mulcahy D, Husain S, Zalaz G. Ischemia during ambulatory monitoring as a prognostic indicator in patients with stable 838 Clinical Investigations

coronary artery disease. JAMA 1997; 277:318 324 10 Sharaf BL, Williams DO, Miele NJ. A detailed angiographic analysis of patients with ambulatory electrocardiographic ischemia: results from the asymptomatic cardiac ischemia pilot (ACIP) study angiographic core laboratory. J Am Coll Cardiol 1997; 79:78 84 11 Narins CR, Zaretra W, Moss AJ, et al. Clinical implications of silent versus symptomatic exercise-induced myocardial ischemia in patients with stable coronary disease. J Am Coll Cardiol 1997; 29:756 763 12 Deedwania P, Carbajal E. Silent ischemia during daily life is an independent predictor of mortality in stable angina. Circulation 1990; 81:748 756 13 Deedwania P, Carbajal E. Usefulness of ambulatory silent myocardial ischemia added to the prognostic value of exercise test parameters in predicting risk of cardiac death in patients with stable angina pectoris and exercise induced myocardial ischemia. Am J Cardiol 1991; 68:1279 1286 14 Rocco M, Nabel E, Campbell S. Prognostic importance of myocardial ischemia detected by ambulatory monitoring in patients with stable coronary artery disease. Circulation 1988; 78:877 884 15 Yeung A, Barry J, Orav J. Effect of asymptomatic ischemia on long-term prognosis in chronic stable coronary disease. Circulation 1991; 83:1598 1604 16 Kennedy HL. Importance of the standard electrocardiogram in ambulatory (Holter) electrocardiography. Am Heart J 1992; 123:1660 1677 17 Crawford MH, Mendoza CA, O Rourke RA, et al. Limitations of continuous ambulatory electrocardiogram monitoring for detecting coronary artery disease. Ann Intern Med 1978; 89:1 5 18 Tzivoni D, Gavish A, Zin D, et al. Prognostic significance of ischemic episodes in patients with previous myocardial infarction. Am J Cardiol 1988; 62:661 664 19 Raby KE, Goldman L, Cook EF, et al. Long-term prognosis of myocardial ischemia detected by Holter monitoring in peripheral vascular disease. Am J Cardiol 1990; 66:1309 1313 20 de Marchena E, Asch J, Martinez J, et al. Usefulness of persistent silent myocardial ischemia in predicting a high cardiac event rate in men with medically controlled, stable angina pectoris. Am J Cardiol 1994; 73:390 392 21 Madjlessi-Simon T, Mary-Krause M, Fillette F, et al. Persistent transient myocardial ischemia despite beta-adrenergic blockade predicts a higher risk of adverse cardiac events in patients with coronary artery disease. J Am Coll Cardiol 1996; 27:1586 1591 22 Deedwania PC, Carbajal EV. Silent myocardial ischemia: a clinical perspective. Arch Intern Med 1991; 151:2373 2382 23 Deedwania PC, Carbajal EV. Prevalence and patterns of silent myocardial ischemia during daily life in stable angina patients receiving conventional antianginal drug therapy. Am J Cardiol 1990; 65:1090 1096 24 Cohn PF. Silent myocardial ischemia: dimensions of the problem in patients with and without angina. Am J Med 1986; 80:3 8 25 Gottlieb SO, Weisfeldt ML, Ouyang P, et al. Silent ischemia as a marker for early unfavorable outcomes in patients with unstable angina. N Engl J Med 1986; 314:1214 1219 26 Currie P, Ashby D, Saltissi S. Prognostic significance of transient myocardial ischemia on ambulatory monitoring after acute myocardial infarction. Am J Cardiol 1993; 71:773 777 27 Gottlieb SO, Gottlieb SH, Achuff SC, et al. Silent ischemia on Holter monitoring predicts mortality in high-risk postinfarction patients. JAMA 1988; 259:1030 1035 28 Deedwania PC. Asymptomatic ischemia during predischarge Holter monitoring predicts poor prognosis in the postinfarction period. Am J Cardiol 1993; 71:859 861 29 Gill JB, Cairns JA, Roberts RS, et al. Prognostic importance of myocardial ischemia detected by ambulatory monitoring early after acute myocardial infarction. N Engl J Med 1996; 334:65 70 CHEST / 120 / 3/ SEPTEMBER, 2001 839