Quantitative T-wave analysis predicts 1 year prognosis and benefit from early invasive treatment in the FRISC II study population

Transcription

1 European Heart Journal (2005) 26, doi: /eurheartj/ehi026 Clinical research Quantitative T-wave analysis predicts 1 year prognosis and benefit from early invasive treatment in the FRISC II study population Michael D. Jacobsen 1 *, Galen S. Wagner 2, Lene Holmvang 1, Frederic Kontny 3, Lars Wallentin 4, Steen Husted 5, Eva Swahn 6, Elisabeth Ståhle 7, Rolf Steffensen 8, and Peter Clemmensen 1 1 The Heart Center, Department of Medicine B, H:S Rigshospitalet, Copenhagen University Hospital, Blegdamsvej 9, DK-2100, Copenhagen, Denmark 2 Duke Clinical Research Institute, Durham, NC, USA 3 The Heart and Lung Center, Ulleval University Hospital, Oslo, Norway 4 Department of Cardiology, Uppsala University Hospital, Uppsala, Sweden 5 Department of Medicine and Cardiology A, Århus University Hospital, Århus, Denmark 6 Department of Cardiology, Heart Center, University Hospital, Linköping, Sweden 7 Department of Thoracic and Cardiovascular Surgery, University Hospital, Uppsala, Sweden 8 Department of Medicine B, Hilleroed Hospital, Hilleroed, Denmark Received 31 October 2003; revised 16 September 2004; accepted 30 September 2004, online publish-ahead-of-print 30 November 2004 See page 103 for the editorial comment on this article (doi: /eurheartj/ehi072) KEYWORDS Acute coronary syndrome; Unstable angina pectoris; Non-ST elevation myocardial infarction Aims To investigate the prognostic value of T-wave abnormalities in patients with non- ST-segment elevation acute coronary syndromes (NSTE-ACS), and whether such ECG changes may predict benefit from an early coronary angiography. Although STsegment changes are considered the most important ECG feature in NSTE-ACS, T-wave abnormalities are the most common ECG finding. We hypothesize that a new quantitative approach to T-wave analysis could improve the prognostic value of this ECG abnormality. Methods and results Quantitative T-wave analysis was performed on the admission ECG in 1609 patients with NSTE-ACS. Nine different categories of T-wave abnormality were analysed for their prognostic value concerning clinical outcome in patients not randomized to early coronary angiography. Also, the presence of one category (i.e. T-wave abnormality in 6 leads) was analysed for its predictive value concerning benefit from early coronary angiography. The combined study endpoint was death or myocardial infarction at 1 year follow-up. Patients with 6 leads with abnormal T-waves and concomitant ST-segment depression had a higher risk when not receiving early coronary angiography (24 vs. 12%, respectively; P ¼ 0.003), but could be brought to the same level of risk as the remaining patients with this treatment. For noninvasively treated patients five different categories of T-wave abnormality were significantly associated with an adverse outcome. Conclusion New quantitative T-wave analysis of the admission ECG gives additional predictive information concerning clinical outcome and identifies patients who benefit from early coronary angiography. * Corresponding author. Department of Medicine B, Hilleroed Hospital, DK-3400 Hilleroed, Denmark. Tel: þ ; fax: þ address: mdiloujacobsen@dadlnet.dk European Heart Journal vol. 26 no. 2 & The European Society of Cardiology 2004; all rights reserved.

2 Quantitative T-wave analysis predicts prognosis 113 Introduction Early coronary angiography (CAG) has recently been established as standard therapy in high-risk patients with non-st-elevation acute coronary syndromes (NSTE-ACS). 1,2 Several clinical and laboratory variables, of which troponins and ST-segment changes on ECG are the most important, are included in the risk stratification. However, we recently found that a quantitative T-wave analysis of the admission ECG based on a newly designed set of normal amplitude ranges for the T-wave could predict an adverse 30 day prognosis in a non-invasively treated NSTE-ACS population. 3 The objective of the present study was to assess whether quantitative T-wave analysis (in patients with or without ST-segment depression) may also provide information on which patients may benefit from early CAG. Furthermore, we sought, in an independent and larger population, to confirm our earlier findings on the predictive value of such analysis concerning clinical outcome in non-invasively treated NSTE-ACS patients. Material and methods Study population This study is a sub-study of the Fragmin and Fast Revascularisation during InStability in Coronary artery disease (FRISC II) study. A total of 2457 patients with NSTE-ACS were included between 1996 and 1998 in 58 centres in Norway, Sweden, and Denmark. Patients were randomized to one of the four following treatment groups as soon as possible and within 72 h: early CAG and long-term dalteparin, early CAG and placebo, non-invasive treatment and long-term dalteparin, and non-invasive treatment and placebo. CAG was to be performed within 7 days after hospital admission. Non-invasive treatment included CAG only when the patient had refractory or recurrent symptoms or a positive exercise test prior to discharge. The study design and results have been described in detail elsewhere. 1,2 Written and oral information was given, and witnessed consent was obtained before inclusion. The main study was approved by the ethics committees of all participating university hospitals. All procedures followed the Declaration of Helsinki. Inclusion criteria were: men.40 years old or post-menopausal women, last episode of angina,48 h before start of dalteparin treatment, and acute ischaemic heart disease verified by cardiac biochemical markers, ST-segment depression, or T-wave inversion 0.1 mv in at least two contiguous leads without the presence of concomitant Q-waves in these leads. Primary main-study exclusion criteria were: age.75 years, indication for thrombolysis, increased risk of bleeding and angioplasty within the last 6 months, or other conditions that made randomization to early revascularization inappropriate. ECG sub-study exclusion criteria were: left ventricular hypertrophy (by the Sokolow Lyon, 4 the Cornell Voltage, 5 or the Romhilt Estes criteria 6 ), right ventricular hypertrophy (by the Butler Leggett criterion 7 ), right and left bundle branch block, atrial flutter and fibrillation, anterior and posterior hemi-block, paced rhythm, Wolff Parkinson White syndrome, low voltage, poor ECG quality, ventricular tachycardia, or heart rate.110 beats per minute. Thus, a total of 1609 patients were included in this sub-study. Of these, 783 (48.7%) were randomized to early CAG and 826 (51.3%) to non-invasive therapy. ECG analyses The admission ECGs were analysed at the ECG core lab at The Heart Center, Department of Medicine B, H: S, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark. ST-segment deviation and T-wave amplitude were measured to the nearest 0.05 mv in all leads except the avr lead. ST-segment deviation was defined as 0.05 mv in any two contiguous leads. Criteria for T-wave abnormality were based on a newly developed set of normal amplitude ranges for the T-wave differentiated by four different age groups (,30, 30 39, 40 49, and 50 years), gender, the 12 ECG leads, and three different QRS axis intervals in the frontal plane (290 to 0, 0 60, ). 3 The nine T-wave abnormality categories developed in a recent study of the thrombin inhibition in myocardial ischaemia (TRIM) population that did not consider T-wave axis were tested in the present study. These are based on ECG lead groups, the number of leads with abnormal T-waves, and the summation of T-wave amplitude abnormalities. The five categories from the TRIM population with a significant relation to outcome were tested in combination. In the TRIM study there was no randomization concerning the interventions, therefore, in the present study the T-wave abnormality criteria were analysed in the 826 patients randomized to non-invasive therapy. The total FRISC II sub-study population was used in the assessment of the relationship between abnormal T-waves and patient outcome with the conservative vs. invasive treatment strategy. Study endpoints The study endpoint was a composite of death or myocardial infarction at 1 year (only counted once per patient). Myocardial infarction was defined by the presence of two or three of the following criteria: typical chest pain, a diagnostic ECG, and elevation of biochemical markers of infarction according to the following definitions: (i) non-procedure-related myocardial infarction, creatine kinase-mb (CK-MB) mass higher than the local limit at one occasion, catalytic activity of CK, CK-B, or CK-MB higher than the local limit at two subsequent measurements or higher than double the local limit at one occasion; (ii) in relation to percutaneous transluminal coronary angioplasty (PTCA), CK-MB mass 1.5 times the local limit at one measurement, catalytic activity of CK, CK-B, or CK-MB three times the upper limit at one occasion or 1.5 times the local limit at two subsequent measurements. In relation to coronary artery bypass surgery only new Q-waves indicated a myocardial infarction. All cases of death or myocardial infarction were adjudicated by an independent endpoint committee. Statistical methods Categorical data are presented by number (%). Continuous variables are presented as medians (interquartile range). All analyses were performed using the STATISTICA, Statsoft program (Tulsa, OK, USA). Cut-off values for T-wave abnormality were identical to those from the recent TRIM study. Continuous variables were compared by the Mann Whitney U test. Comparisons of categorical characteristics was by Yates s corrected x 2 P-values. Both univariable and multivariable analyses were applied to examine the association between T-wave abnormalities and myocardial infarction or death at 1 year. Univariable analysis was performed by log rank survival analysis using two-sided tests. Multivariable analysis was performed using backwards stepwise Cox regression. The following variables, with P, 0.20 in the univariable analysis, were included in the

3 114 M.D. Jacobsen et al. stepwise analysis: previous myocardial infarction, IDDM, beta-blocker treatment, ST-segment depression, and T-wave abnormality categories 1, 2, 3, 6, 7, and 9. No violations of the proportional hazards assumption were detected by survival analysis comparing patients with and without T-wave abnormality. Relative risks were presented with 95% confidence intervals. Results Baseline characteristics and patient outcomes A total of 1534 patients (95.3%) had follow-up data at 1 year. The censoring of data for the remaining patients was not taken into account. The median follow-up time (for uncensored patients) was 378 days; interquartile range 27 days. The baseline characteristics are presented in Table 1. The early CAG and the non-invasive groups were significantly different with regard to gender. Of note, the FRISC II population was a relatively high-risk population with a large proportion presenting with ST-segment depression and positive troponin T. Prognostic value of T-wave abnormality Tables 2 and 3 show the relation of the T-wave abnormality categories to the study endpoints in the 826 patients randomized to non-invasive therapy differentiated to patients without ST-segment depression (Table 2, n ¼ 382), and to patients with concomitant ST-segment depression (Table 3, n ¼ 444; findings for Table 3 must be considered in the light of the many comparisons of data). When no ST-segment depression was present, the T-wave abnormality categories had no significant prognostic value. However, in patients with ST-segment depression, five different categories of concomitantly Table 1 The baseline characteristics of the FRISC II population (1 year follow-up) present T-wave abnormality significantly predicted an adverse outcome compared with patients with only ST-segment depression. The relative risks for the categories of 6 leads with abnormal T-waves [1.93; confidence interval (CI) ], S T-wave amplitude abnormality (1.82; CI ), an abnormal T-wave in any of leads V 4 6 (2.18; CI ), and S T- wave amplitude abnormality in leads V 4 6 (2.02; CI ) were greater than the relative risk in the presence of type 1 diabetes (1.34; CI ). Also, the algorithm of combined abnormality categories (1, 2, 4, 6, 7), significantly identified high-risk patients. Table 4 shows the results of a multivariable analysis regarding the prediction of death or myocardial infarction. Among the electrocardiographic findings, ST-segment depression was the strongest predictor but the total number of abnormal T-waves in 6 leads was close to being equally significant. Predictive value of T-wave abnormality on patient outcome in conservative vs. early invasive treatment For this analysis, only T-wave abnormality category 1 was chosen, since it had the strongest P-value in univariable analysis and was the only significant category in the multivariable analysis. Among patients with 6 leads with abnormal T-waves and concomitant ST-segment depression, conservative treatment was, compared with early CAG, associated with significantly higher risk for adverse clinical outcome at 1 year (24 vs. 12%, respectively; P ¼ 0.003). When given invasive treatment, the risk approximated that of patients with,6 leads with abnormal T-waves receiving non-invasive treatment (Figure 1 ). This benefit of treatment could not be Total Non-invasive Invasive P n Median age (interquartile range) 65 (57 71) 64 (57 71) 65 (56 72) Men 1072 (67) 531 (64) 541 (69) Insulin-dependent diabetes mellitus 75 (5) 33 (4) 42 (5) Hypertension 454 (28) 242 (29) 212 (27) Previous myocardial infarction 335 (21) 168 (20) 167 (21) Previous angioplasty, bypass, or both 57 (4) 22 (3) 35 (4) Beta-blocker at admission 502 (31) 261 (32) 241 (31) Digoxin at admission 21 (1) 11 (1) 10 (1) Admission ischemia T-wave inversion a 904 (56) 646 (78) 601 (77) ST-segment depression 847 (53) 444 (54) 403 (51) TnT positive (non-st-segment elevation AMI) b 871 (54) 444 (54) 427 (55) Follow-up New acute myocardial infarction 160 (10) 97 (12) 63 (8) Death, all causes 45 (3) 28 (3) 17 (2) Combined endpoint (death or new acute myocardial infraction 189 (12) 114 (14) 75 (10) Difference between invasive- and non-invasive groups. Yates s corrected x 2. Figures in brackets are percentages. a Main study inclusion criterion. b Troponin T cut-off value: 0.1 mg/l.

4 Quantitative T-wave analysis predicts prognosis 115 Table 2 Prognostic significance of various categories of T-wave abnormality concerning death or acute myocardial infarction at 1-year follow-up in the 382 patients without ST-segment depression randomized to non-invasive therapy Abnormality category Cut-off n (%) P All leads 1 Total number of leads with abnormal T-waves (34) S T-wave amplitude abnormality 5.0 mm 177 (46) Largest single T-wave amplitude abnormality 2.0 mm 180 (47) Leads V An abnormal T-wave in any of leads V (57) S T-wave amplitude abnormality in leads V mm 185 (48) Leads V An abnormal T-wave in any of leads V (60) S T-wave amplitude abnormality in leads V mm 152 (40) Leads II, avf, III 8 An abnormal T-wave in any of leads II, avf, III 235 (62) S T-wave amplitude abnormality in leads II, avf, III 0.5 mm 169 (44) Combined T-wave categories 1, 2, 4, 6, and (82) Univariable log rank survival analysis. Table 3 Prognostic significance of various categories of T-wave abnormality concerning death or acute myocardial infarction at 1-year follow-up in the 444 patients with concomitant ST-segment depression randomized to non-invasive therapy Abnormality category Cut-off n (%) P RR a CI (95%) All leads 1 Total number of leads with (46) abnormal T-waves 2 S T-wave amplitude abnormality 5.0 mm 278 (63) Largest single T-wave amplitude abnormality 2.0 mm 268 (60) Leads V An abnormal T-wave in any of leads V S T-wave amplitude abnormality in leads V 1 3 Leads V An abnormal T-wave in any of leads V S T-wave amplitude abnormality in leads V 4 6 Leads II, avf, III 8 An abnormal T-wave in any 320 (72) of leads II, avf, III 9 S T-wave amplitude abnormality in leads 0.5 mm 258 (58) II, avf, III Combined T-wave categories 1, 2, 4, 6, and (90) Univariable log rank survival analysis. a Relative risk. found in patients with 6 leads with abnormal T-waves and an isoelectric ST-segment (Figure 2 ). Discussion This FRISC II sub-study elaborated on the findings of the recent TRIM sub-study, 3 and found five of nine recently developed categories for abnormal T-waves to be significantly related to a worse prognosis. This finding is almost identical to that of the TRIM population, although one category (largest single T-wave abnormality) was significant in this study but not in the TRIM study. In the multivariable analysis, among the electrocardiographic findings, ST-segment depression was found to be the strongest single independent predictor of outcome, but the category of 6 leads with abnormal T-waves was almost equally significant and the relative risks for the two variables were similar. In the TRIM study, in patients without ST-segment depression categories 1 and 6

5 116 M.D. Jacobsen et al. showed a trend towards significance. In the present study, no relation between T-wave abnormality and outcome could be found for patients with a normal ST-segment. The reason for this discrepancy could of course be a type 1 error due to a relatively small sample size in that TRIM sub-group; it could perhaps also be found in the difference between the two populations. The TRIM population was a more chronic population (45% previous AMI, 8% ST-segment depression) and the FRISC II population a more acute population (21% previous AMI, 53% ST-segment depression); P, for both variables. Finally, in this study, quantitative T-wave analysis identified patients who benefit from an early CAG. Table 4 Multivariable analysis of baseline predictors and study variables concerning death or acute myocardial infarction within 1 year, in 826 patients with NSTE-ACS randomized to non-invasive therapy Variables P RR a CI (95%) Previous myocardial infarction ST-segment depression Total number of leads with abnormal T-waves (6) Beta-blocker at admission S T-wave amplitude abnormality in leads II, avf, III Type 1 diabetes a Relative risk. In the FRISC II study, patients with normal ECGs and normal biochemical markers were excluded. The prognostic value of an abnormal T-wave in this setting therefore applies to patients with a high probability of having a true unstable coronary syndrome. The true prognostic value of abnormal T-waves in patients presenting with acute chest pain may therefore be even greater. Previous studies show that T-wave changes are far more common than ST-segment changes in patients with NSTE-ACS 3 (A.M. Kucia et al. Improved detection of ischaemia by the addition of continuous T-wave monitoring rather than continuous ST-segment monitoring alone in patients with unstable angina pectoris. Manuscript in preparation). It is the authors impression that cardiologists do look to the ECG T-wave changes and their dynamics in patients with acute coronary syndromes. Disappointingly, when tested in retrospect, T-wave changes have had little prognostic value in several large randomized clinical trials None of these used a quantitative approach to T-wave analysis, nor did they correct for ageand gender-adjusted normal values. In a global utilization of streptokinase and t-pa for occluded coronary arteries (GUSTO)-IIa sub-study, Ohman et al. 11 pooled patients with T-wave inversion and isoelectric ST-segment with patients with normal tracings. Not surprisingly, significantly fewer of these patients died within 30 days compared with patients with ST-segment depression and with ST-segment elevation (1.2, 8.0, and 7.4%, respectively). However, the findings of Savonitto et al. 12 in the GUSTO-IIb study suggest that patients with ST-depression and isolated T-wave inversion have equal degrees of coronary artery disease: elevated CK at the Figure 1 Kaplan Meier plot of the event-free proportion (freedom from death and acute myocardial infarction) of patients with ST-segment depression and randomized to early coronary angiography divided by number of leads with abnormal T-waves (n ¼ 84). Early coronary angiography.

6 Quantitative T-wave analysis predicts prognosis 117 Figure 2 Kaplan Meier plot of the event-free proportion (freedom from death and acute myocardial infarction) of patients without ST-segment depression and randomized to early coronary angiography divided by number of leads with abnormal T-waves (n ¼ 762). Early coronary angiography. time of admission occurred with identical frequency (12%) in these groups of patients, and at 30-day follow-up, the need for any revascularization was also similar (36 and 37%, respectively). Newby et al. 13 tested the prognostic value of a combined multi-marker of myoglobin, CK-MB, and troponin I in 1005 patients with chest pain. 13 Of patients with T-wave inversion, 13.1% were positive for the combined marker and 9.2% were negative. For patients with ST-segment elevation, the numbers were 12.4 and 10.5%, respectively, and for patients with ST-segment depression 7.6 and 4.9%, respectively (no P-value calculated). The results suggest that similar ratios of patients with any of these three different ECG readings have myocardial damage. The thrombolysis in myocardial infarction (TIMI) risk score offered an easily applicable model for early risk stratification. 14 It was based on baseline characteristics, coronary artery stenosis detected in a prior coronary angiography, biochemical markers, and ST-segment deviation. Seven different categories were found to carry independent prognostic information, but only the ST-segment was considered in the ECG. Quantitative T-wave analysis might further refine this and other clinical risk scores. Given the necessity for the use of a table of normal T-wave amplitudes, however, quantitative T-wave analysis is ideally suited for incorporation in computer software for automatic continuous analysis. Recently, T-wave fluctuations in continuous monitoring were found to be significantly correlated with significant coronary artery stenosis (A.M. Kucia et al., in preparation). It is a weakness of this study that there is a significant gender difference between the invasively and the non-invasively treated groups. Conclusion With the results of the FRISC II main study, the early invasive treatment strategy of patients with non-st-segment elevation acute coronary syndromes has been further promoted. The present sub-study supports the findings of the recent TRIM sub-study, 3 and emphasizes the importance of quantitative T-wave analysis to supplement ST-segment analysis of the admission ECG for early risk stratification and thereby a rational use of invasive-cardiology facilities. Acknowledgements The authors wish to thank Charles Maynard of the Department of Health Services, University of Washington, Seattle, WA, USA for statistical assistance. The FRISC II study was financially supported by Pharmacia. The ECG core lab at the Heart Center, Rigshospitalet was supported through a grant from the Danish Heart Foundation. M.D.J. was supported by a grant from The Danish Medical Society. References 1. Invasive compared with non-invasive treatment in unstable coronaryartery disease: FRISC II prospective randomised multicentre study. FRagmin and Fast Revascularisation during InStability in Coronary artery disease Investigators. Lancet 1999;354: Wallentin L, Lagerqvist B, Husted S, Kontny F, Stahle E, Swahn E. Outcome at 1 year after an invasive compared with a non-invasive strategy in unstable coronary-artery disease: the FRISC II invasive

7 118 M.D. Jacobsen et al. randomised trial. FRISC II Investigators. Fast Revascularisation during Instability in Coronary artery disease. Lancet 2000;356: Jacobsen MD, Wagner GS, Holmvang L, Macfarlane PW, Näslund U, Grande P, Clemmensen P on behalf of the TRIM investigators. Clinical significance of abnormal T waves in patients with non STsegment elevation acute coronary syndromes. Am J Cardiol 2001;88: Sokolow M, Lyon TP. The ventricular complex in left ventricular hypertrophy as obtained by unipolar precordial and limb leads. Am Heart J 1949;40: Casale PN, Devereux RB, Kligfield P. Electrocardiographic detection of left ventricular hypertrophy development and prospective validation of improved criteria. J Am Coll Cardiol 1985;6: Reichek N, Devereux RB. Left ventricular hypertrophy: relationship of anatomic, echocardiographic and electrocardiographic findings. Circulation 1981;63: Butler PM, Leggett SI, Howe CM, Freye CJ, Hindman NB, Wagner GS. Identification of electrocardiographic criteria for diagnosis of right ventricular hypertrophy clinical trial: results with and without the end-point committee s final decision on end-points. Eur Heart J 1999;20: Hyde TA, French JK, Wong C, Staznicky IT, Whitlock RM, White HD. Four-year survival of patients with acute coronary syndromes without ST-segment elevation and prognostic significance of 0.5-mm ST-segment depression. Am J Cardiol 1999;84: Patel DJ, Holdright DR, Knight CJ, Mulcahy D, Thakrar B, Wright C. Early continuous ST segment monitoring in unstable angina prognostic value additional to the clinical characteristics and the admission electrocardiogram. Heart 1996;75: Holmvang L, Clemmensen P, Wagner GS, Grande P. Admission standard electrocardiogram for early risk stratification in patients with unstable coronary artery disease not eligible for acute revascularization therapy: a TRIM substudy. ThRombin Inhibition in Myocardial infarction. Am Heart J 1999;137: Ohman EM, Armstrong PW, Christenson RH, Granger CB, Katus HA, Hamm CW, O Hanesian MA, Wagner GS, Kleiman NS, Harrell FE Jr, Califf RM, Topol EJ. Cardiac troponin T levels for risk stratification in acute myocardial ischemia. GUSTO IIa investigators. N Engl J Med 1996;335: Savonitto S, Ardissino D, Granger CB, Morando G, Prando MD, Mafrici A, Cavallini C, Melandri G, Thompson TD, Vahanian A, Ohman EM, Califf RM, Van de Werf F, Topol EJ. Prognostic value of the admission electrocardiogram in acute coronary syndromes. JAMA 1999;281: Newby LK, Storrow AB, Gibler WB, Garvey JL, Tucker JF, Kaplan AL, Schreiber DH, Tuttle RH, McNulty SE, Ohman EM. Bedside multimarker testing for risk stratification in chest pain units. Circulation 2001; 103: Antman EM, Cohen M, Bernink PJ, McCabe CH, Horacek T, Papuchis G, Mautner B, Corbalan R, Radley D, Braunwald E. The TIMI risk score for unstable angina/non-st elevation MI: a method for prognostication and therapeutic decision making. JAMA 2000;284: