International Journal of Cardiology 61 (1997) 239–245
Effect of ST segment measurement point on performance of exercise ECG analysis a, b ¨ ¨ ¨ Turjanmaa c , Kari Niemela¨ d , Jaakko Malmivuo a Rami Lehtinen *, Harri Sievanen , Vaino a
Ragnar Granit Institute, Tampere University of Technology, P.O. Box 692, FIN-33101 Tampere, Finland b UKK Institute for Health Promotion Research, Tampere, Finland c Department of Clinical Physiology, Tampere University Hospital and Medical School, University of Tampere, Tampere, Finland d Division of Cardiology, Department of Internal Medicine, Tampere University Hospital, Tampere, Finland Received 26 March 1997; received in revised form 26 June 1997; accepted 26 June 1997
Abstract To evaluate the effect of ST-segment measurement point on diagnostic performance of the ST-segment / heart rate (ST / HR) hysteresis, the ST / HR index, and the end-exercise ST-segment depression in the detection of coronary artery disease, we analysed the exercise electrocardiograms of 347 patients using ST-segment depression measured at 0, 20, 40, 60 and 80 ms after the J-point. Of these patients, 127 had and 13 had no significant coronary artery disease according to angiography, 18 had no myocardial perfusion defect according to technetium-99m sestamibi single-photon emission computed tomography, and 189 were clinically ‘normal’ having low likelihood of coronary artery disease. Comparison of areas under the receiver operating characteristic curves showed that the discriminative capacity of the above diagnostic variables improved systematically up to the ST-segment measurement point of 60 ms after the J-point. As compared to analysis at the J-point (0 ms), the areas based on the 60-ms point were 89 vs. 84% ( p50.0001) for the ST / HR hysteresis, 83 vs. 76% ( p,0.0001) for the ST / HR index, and 76 vs. 61% ( p,0.0001) for the end-exercise ST depression. These findings suggest that the ST-segment measurement at 60 ms after the J-point is the most reasonable point of choice in terms of discriminative capacity of both the simple and the heart rate-adjusted indices of ST depression. Moreover, the ST / HR hysteresis had the best discriminative capacity independently of the ST-segment measurement point, the observation thus giving further support to clinical utility of this new method in the detection of coronary artery disease. 1997 Elsevier Science Ireland Ltd Keywords: Exercise ECG; Computer analysis; Coronary artery disease; Noninvasive diagnosis; Medical informatics; Receiver operating characteristic analysis
1. Introduction Analysis of the exercise-induced ST-segment response during exercise ECG test is the most widely used noninvasive method for the detection of coronary artery disease. The ST-segment depression / heart rate (ST / HR) analysis during exercise phase, *Corresponding author. Tel.: 1358 3 3652524; fax: 1358 3 3652162; e-mail:
[email protected].
i.e., the ST / HR slope [1] and the ST / HR index [2], as well as the novel continuous diagnostic variables, the ST / HR hysteresis [3–5] and the recovery loop index [6] which combine the ST / HR analysis during both the exercise and postexercise recovery phases of the exercise ECG test, have been shown to improve the detection of coronary artery disease. However, further refinement and evaluation of the ST / HR analysis requires adequate knowledge on the effect of ST-segment measurement point on the diagnostic
0167-5273 / 97 / $17.00 1997 Elsevier Science Ireland Ltd All rights reserved. PII S0167-5273( 97 )00157-5
240
R. Lehtinen et al. / International Journal of Cardiology 61 (1997) 239 – 245
performance of the above diagnostic variables. Even though it is obvious that the magnitude of the STsegment depression measured at given time after the J-point is influenced by the ST-segment configuration and by the shortening of the QT-interval at high heart rates, there exists no standard for the optimal time after J-point at which to measure the ST-segment depression during the exercise test. Recently, the ST-segment measurement points at 60 ms after the J-point (J160) and at the J-point (J10) have been compared in connection with the debate on the improved diagnostic performance of the ST / HR slope [7,8] and ST / HR index [7–10] in the detection of coronary artery disease but in a contradictory fashion. As compared to the J160, the J10 has been shown to decrease [7,8], be equal to [9], or increase [10] the discriminative capacity of the ST / HR index. To elaborate the issue of selecting an appropriate measurement point, the objective of this study was to evaluate the effect of different J-point fixed STsegment measurement points on the performance of the ST / HR hysteresis, the ST / HR index, and the end-exercise ST-segment depression in the detection of coronary artery disease.
2. Methods
2.1. Study population Exercise electrocardiographic data from a total of 347 patients, referred by a physician for a routine clinical exercise ECG test due to symptoms or
clinical signs suggesting cardiovascular or pulmonary disease or limitation in working capacity were taken from our previous study [3] and reanalysed for the study. The description of the study population is given in Table 1. None of the patients had a left or right bundle branch block pattern on the resting ECG or recent (,8 weeks) myocardial infarction. Of the 347 patients, 140 were investigated by coronary angiography within 180 days of the exercise test according to which 127 had and 13 had not significant ($50%) coronary artery stenosis. In addition, the study population included 18 patients who had no myocardial perfusion defect according to technetium99m sestamibi single-photon emission computed tomography and 189 clinically ‘normal’ patients with respect to cardiac diseases who had no history of any cardiac disease, had a normal resting ECG, and had no anginal chest pain or cardiac medication.
2.2. Exercise ECG test All subjects were tested on a bicycle ergometer. The exercise protocols were individualised to some extent depending on the patient’s physical condition. The protocol followed a standard clinical routine with an initial workload of 40 W for women and 50 W for men and an increment of 40 or 50 W every 4 min for women and men, respectively [11]. The ECG recordings were made with a commercial ECG recording system (SYSTEM II EXES; SiemensElema, Solna, Sweden). The ST-segment and heart rate data were stored for further processing and analysis. The ECG lead system used in the exercise
Table 1 Study population Characteristic
Age (years) Sex (male / female) Medication (1 / 2) b-Blockers Calcium blockers Nitrates Anginal chest pain (1 / 2) MaxHR (beats / min)
Angiographically
Clinical reference group
proven CAD (n5127)
Total (n5220)
No CAD by angiography (n513)
No Ischemia by MIBI SPECT (n518)
Clinically normal (n5189)
5568 101 / 26
48612 113 / 107
5265 4/9
5069 9/9
47612 100 / 89
104 / 23 46 / 81 86 / 41 60 / 67 125621
12 / 208 4 / 216 5 / 215 3 / 217 162619
9/4 4/9 5/8 3 / 10 145623
3 / 15 0 / 18 0 / 18 0 / 18 162617
0 / 189 0 / 189 0 / 189 0 / 189 164619
Data taken from our previous study [3], continuous data given as mean6SD. CAD, coronary artery disease; MaxHR, maximum heart rate achieved; MIBI SPECT, technetium-99m sestamibi single-photon emission computed tomography.
R. Lehtinen et al. / International Journal of Cardiology 61 (1997) 239 – 245
test was the Mason-Likar modification of the standard 12-lead system. Exercise tests were sign- or symptom-limited maximal tests using recommended criteria [11] for termination; exhaustion or chest pain being the most common reasons for termination.
2.3. The exercise ECG variables Through performing a computer analysis of the stored ST-segment and heart rate data, the maximum values of the ST / HR hysteresis, ST / HR index, and end-exercise ST depression were determined from the 12-lead system (leads aVL, aVR and V1 excluded) for each patient without knowledge of his / her cardiac status [12]. The ST-segment depressions used in constructing all the above diagnostic variables were measured to the nearest 10 mV at 0, 20, 40, 60, and 80 ms after the J-point. Of note, an ST-segment elevation was expressed as a negative ST-segment depression. The determination of the ST / HR hysteresis for each lead was performed identical to the recent reports [3,12] as illustrated in Fig. 1. In short, the ST / HR diagram containing the pairs of the ST depression and heart rate was determined by measuring the given data immediately before starting the
241
exercise with the patient sitting on the bicycle, at the end of each minute of exercise, at the end-exercise, and at the end of each of the first three consecutive minutes of postexercise recovery. A continuous piecewise linear function was obtained by connecting the consecutive ST / HR data pairs of the exercise phase (closed circles in Fig. 1) with lines. Similarly, a continuous piecewise linear function for the postexercise recovery phase was constructed by connecting the consecutive ST / HR data pairs of the first 3 min of recovery phase (open circles in Fig. 1) starting from the ST / HR data pair at the end-exercise. Then the difference between the above curves was integrated over the heart rate from the minimum heart rate of recovery (HR rec in Fig. 1) to the maximum heart rate (HR exe in Fig. 1). Finally the ST / HR hysteresis was obtained by dividing the integrated net difference by the heart rate difference of the integration interval (DHR rec in Fig. 1) in order to normalize the result with respect to the postexercise heart rate decrement. The ST / HR index was calculated as the gradient between the ST / HR pairs at the start-exercise and at the end-exercise as suggested by Detrano et al. [2].
2.4. Coronary angiography The selective coronary angiography was performed with the Judkins technique. Each coronary artery was imaged in multiple views in all cases. A degree of stenosis was defined as the greatest percentage reduction of luminal diameter in any view compared with the nearest normal segment without knowledge of the exercise ECG data. Coronary artery disease was considered significant when 50% or more luminal narrowing was observed at least in one major coronary artery: in the left main, in the left anterior descending, in the left circumflex or in the right coronary artery.
2.5. Technetium-99 m sestamibi myocardial imaging
Fig. 1. Determination of ST / HR hysteresis. In this diagram, the ST depression is plotted in an upward direction on the vertical axis, and an ST elevation is expressed as a negative ST depression. See text for details.
The myocardial perfusion imaging followed the standard clinical routine. Abnormalities in the myocardial perfusion were identified as abnormal distribution images. Distribution images were determined by computer analysis of the results obtained from the single-photon emission computed tomog-
242
R. Lehtinen et al. / International Journal of Cardiology 61 (1997) 239 – 245
raphy. The visual interpretation of regional perfusion defects were determined in the septal, anterior, lateral, posterior, inferior and apical regions of the left ventricle.
2.6. Data analysis and statistical methods Mean values of the descriptive variables were reported with the standard deviation (SD). The discriminative capacity of the continuous diagnostic variables, the ST / HR hysteresis, the ST / HR index and the end-exercise ST depression were compared independently of the operating point (i.e. partition value) selection by means of the receiver operating characteristics analysis. The area under the receiver operating characteristic curve represents the discriminative capacity; i.e., the probability that a random pair of patients with and without coronary artery disease will be correctly diagnosed [13]. The differences between the areas under the receiver operating characteristic curves were compared using a computer program [14] (version 2.5, McGill University, Montreal, Canada) for nonparametric analysis of correlated receiver operating characteristic curves [15]. For all comparisons, a value of p,0.05 was used for statistical significance.
3. Results The receiver operating characteristic curves of the ST / HR hysteresis, ST / HR index and end-exercise ST depression, based on ST-segment measurement at 60 ms after the J-point and at the J-point of 0 ms, are given in Fig. 2. Comparison of the areas under the receiver operating characteristic curves showed the superior discriminative capacity of the J160 as opposed to the J10 of the ST / HR hysteresis (89 vs. 84%, p50.0001), the ST / HR index (83 vs. 76%, p,0.0001), and end-exercise ST depression (76 vs. 61%, p,0.0001). The areas under the ROC curves of the diagnostic variables based on ST-segment measurement at 0, 20, 40, 60 and 80 ms after the J-point are illustrated in Fig. 3. The ST / HR hysteresis had significantly the largest areas under the ROC curves no matter on which ST-segment measurement point the diagnostic variables were based, the finding indicating the best discriminative capacity of the
Fig. 2. Receiver operating characteristic curves for the ST / HR hysteresis (upper panel), the ST / HR index (middle panel), and the end-exercise ST depression (lower panel) for the ST-segment measurement points at the J-point (J10) and at 60 ms after the J-point (J160). The curve symbols refer to the partition values of the variables of which some are specified, expressed in mV for the ST / HR hysteresis and end-exercise ST depression, and in mV/ beats / min for the ST / HR index. AUC, area under the receiver operating characteristic curve.
R. Lehtinen et al. / International Journal of Cardiology 61 (1997) 239 – 245
243
4. Discussion
Fig. 3. Areas under the receiver operating characteristic (ROC) curves of the ST / HR hysteresis, ST / HR index and end-exercise ST depression with different ST-segment measurement points.
ST / HR hysteresis independently of the ST-segment measurement point. The diagnostic performance of the above three variables using fixed positive test criteria (i.e., the partition values given in Fig. 2) were clearly different when the J160 was contrasted to the J10. The supplementary data from measurement points 20, 40 and 80 ms after the J-point (Table 2) showed a clear trend, which indicated that the earlier the measurement point, the lower the specificity and the higher the sensitivity.
Table 2 Specificity and sensitivity of the ST / HR hysteresis, ST / HR index, and end-exercise ST depression using the same fixed partition values with different ST-segment measurement points Positive test criterion
J10
J120
ST / HR hysteresis $0.01 mV Specificity (%) 46 65 Sensitivity (%) 91 88 ST / HR index $1.6 mV/ beats / min Specificity (%) 37 60 Sensitivity (%) 84 78 End-exercise ST depression $0.10mV Specificity (%) 33 55 Sensitivity (%) 76 69
J140
J160
J180
77 84
83 80
88 74
72 73
81 72
88 64
75 61
82 55
86 50
Results of the present study demonstrated that the use of the ST-segment measurement point at the J-point rather than at 60 ms after the J-point reduces the discriminative capacity of the ST / HR hysteresis, ST / HR index and end-exercise ST depression in the detection of coronary artery disease. This was corroborated by the supplementary analysis of the STsegment measurement points at 20, 40 and 80 ms after the J-point, which showed that the closer the measurement point to the J-point the poorer the discriminative capacity of the diagnostic variable (Fig. 3). This trend can be explained by the apparent interrelations between the ST-segment amplitude and configuration at a given time point (see below), and between the length of the QT-interval and heart rate. Depending on the ST-segment configuration (in brackets), the later ST-segment measurement point will decrease (upsloping ST-segment), not affect (horizontal ST-segment) or increase (downsloping ST-segment) the magnitude of the measured ST depression. The shortening of the QT-interval with high heart rate makes the later ST-segment measurement points (J160 or J180) coincide with the very beginning of T wave, decreasing the magnitude of the measured ST-segment depression at the same. The upsloping ST-segment configuration [16] and high maximum heart rate achieved [17,18] are more likely to appear in patients without coronary artery disease, thus enhancing the discrimination of those patients from patients with coronary artery disease. A similar trend from J10 to J160 in ST / HR index and endexercise ST depression was also found by Okin et al. [8], except the J180 showed a slightly poorer discriminative capacity than the J160, which finding they stated to be due to a coincidence of J180 measurement point with the very beginning of the T wave in patients with coronary artery disease. An apparent reason why we did not observe this effect was that our patients with coronary artery disease attained slightly lower maximum heart rate than did their patients. The ST-segment measurement point seemed to have a substantial effect on the specificity and sensitivity of the ST / HR hysteresis, ST / HR index and end-exercise ST depression when using a fixed partition value (Table 2). The later the ST-segment
244
R. Lehtinen et al. / International Journal of Cardiology 61 (1997) 239 – 245
measurement point, the higher the specificity and the lower the sensitivity in all the diagnostic variables. For example with the positive test criterion of endexercise ST depression $0.10 mV, the specificity of the J10 was 33% with a sensitivity of 76%, whereas using the J160 the specificity was 82% with a sensitivity of 55%. These results strongly suggest that different partition values should be used for different ST-segment measurement points, when applying the variables in diagnostic decision making. The results of this study support the findings of the Cornell group [7,8], but are not in concordance with the results of the Long Beach group [9,19] nor the Poitiers group [10]. These divergent results may stem from the differences in the populations, which have a well-recognised effect on the test performance [20,21]. The major difference in the previous studies was in the ‘normal’ population in which both the Long Beach [9,19] and the Poitiers groups [10] used only the catheterised patients without coronary artery disease, whereas the Cornell group [7,8] excluded these patients and included only subjects with a low likelihood of coronary artery disease. Defining the ‘normal’ population for evaluation of exercise ECG test is problematic [20]; the catheterised patients without coronary artery disease commonly have important myopathic disease and occasionally have small-vessel coronary disease [22], whereas the subjects whose status is not verified by coronary angiography or isotope single-photon emission computed tomography myocardial imaging are of limited challenge [23]. In the ‘normal’ population of the present study, we included catheterised patients without coronary artery disease, technetium-99m sestamibi single-photon emission computed tomography myocardial imaging verified patients without perfusion defect, and clinically ‘normal’ patients with respect to cardiac diseases who had no history nor evident signs or symptoms of cardiac disease. Although ischemic heart disease cannot be totally excluded from the clinical reference group of the present study, clinical examinations, normal resting ECG, and probabilistic assessment [24] suggested that only a very small fraction of these subjects was likely to have significant ischemic heart disease, and the group could be considered normal with respect to coronary artery disease or myocardial ischemia. The
patients with angiographically proven coronary artery disease of this study represent a population routinely referred for coronary angiography. Collectively, the study population of this study can be considered to represent the majority of the patients, symptomatic or not, referred in clinical practice for screening exercise ECG. It is actually in this population where highperformance diagnostic criteria and tests are required. The ST / HR hysteresis had the best discriminative capacity no matter on which ST-segment measurement point the analysis was based (Fig. 3). Furthermore its discriminative capacity was least affected by the ST-segment measurement point. Probably this beneficial property of the ST / HR hysteresis is due to its ability to integrate the exercise and recovery ST / HR analysis in which the prevailing direction and average magnitude of the hysteresis in ST depression against heart rate is determined [3]. The utility of this approach is also supported by recent results of Herpin et al. [6], who using a recovery loop index, a measure comparable to the ST / HR hysteresis, showed that its discriminative capacity was virtually equal with J120 and J160 (areas under the ROC curves, 83 vs. 84%). In conclusion, the findings of this study suggest that the ST-segment measurement at 60 ms after the J-point is the most reasonable point of choice in terms of discriminative capacity of both the simple and the heart rate-adjusted indexes of ST depression. Moreover, this study showed that the ST / HR hysteresis had the best discriminative capacity independently of the ST-segment measurement point, the observation thus giving further support to the clinical utility of this new method in the detection of coronary artery disease.
Acknowledgments We thank Jukka Niemi, MSc, for his technical support in processing the data. The financial support of the Academy of Finland, Alfred Kordelin Foundation, Ella and Georg Ehrnrooth Foundation, Finnish Cultural Foundation (Fund of Pirkanmaa), Finnish Cultural Foundation, Finnish Foundation for Cardiovascular Research, Ida Montini Foundation, Ragnar Granit Foundation, Tampere Science Foundation, and Wihuri Foundation is gratefully acknowledged.
R. Lehtinen et al. / International Journal of Cardiology 61 (1997) 239 – 245
References [1] Elamin M, Mary D, Smith D, Linden R. Prediction of severity of coronary heart disease using slope of submaximal ST-segment / heart rate relationship. Cardiovasc Res 1980;14(14):681–91. [2] Detrano R, Salcedo E, Passalacqua M, Friis R. Exercise electrocardiographic variables: a critical appraisal. J Am Coll Cardiol 1986;8(4):836–47. ¨ [3] Lehtinen R, Sievanen H, Viik J, Turjanmaa V, Niemela¨ K, Malmivuo J. Accurate detection of coronary artery disease by integrated analysis of the ST-segment depression / heart rate patterns during the exercise and recovery phases of the exercise electrocardiography test. Am J Cardiol 1996;78(9):1002–6. ¨ [4] Lehtinen R, Sievanen H, Viik J, Vuori I, Malmivuo J. Reproducibility of the ST-segment depression / heart rate analysis of the exercise ECG test in asymptomatic middle-aged population. Am J Cardiol 1997;79(10):1414–6. [5] Viik J, Lehtinen R, Turjanmaa V, Niemela¨ K, Malmivuo J. Effect of lead selection on conventional and heart rate adjusted ST-segment analysis in detection of coronary artery disease during exercise testing. Am Heart J 1997:in press. [6] Herpin D, Ferrandis J, Couderq C. Usefulness of a quantitative analysis of the recovery phase patterns of the ST-segment depression in the diagnosis of coronary artery disease. Am J Med 1996;101:592–8. [7] Kligfield P, Okin PM. The academic life of a noninvasive test (letter). Circulation 1990;82(6):2284–5. [8] Okin PM, Bergman G, Kligfield P. Effect of ST segment measurement point on performance of standard and heart rate-adjusted ST segment criteria for the identification of coronary artery disease. Circulation 1991;84(1):57–66. [9] Rodriguez M, Froning J, Froelicher VF. ST0 or ST60. Am Heart J 1993;126(3):752–4. [10] Herpin D, Ferrandis J, Borderon P. Comparison of the diagnostic accuracy of different methods of measurement of heart rate-adjusted ST-segment depression during exercise testing for identification of coronary artery disease. Am J Cardiol 1995;76:1147–51. ¨ [11] Arstila M, Kallio V, Seppanen A, editors. Suositus kliinisen rasituskokeen suorittamisesta ja tulkinnasta (Clinical exercise testing: Standards for procedures and recommendations for the interpreta¨ tion). Turku: Kansanelakelaitos kuntoutustutkimuskeskus, 1984.
245
¨ ¨ [12] Lehtinen R, Vanttinen H, Sievanen H, Malmivuo J. A computer program for comprehensive ST-segment depression / heart rate analysis of the exercise ECG test. Comput Methods Programs Biomed 1996;50(1):63–71. [13] Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143(1):29–36. [14] Vida S. A computer program for non-parametric receiver operating characteristic analysis. Comput Methods Programs Biomed 1993;40:95–101. [15] DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics 1988;44:837–45. [16] Sheffield LT. Upsloping ST segments. Easy to measure, hard to agree upon. Circulation 1991;84(1):426–8. [17] Chin CF, Messenger JC, Greenberg PS, Ellestad MH. Chronotropic incompetence in exercise testing. Clin Cardiol 1979;2:12–8. [18] Brenner SJ, Pashkow FJ, Harvey SA, Marwick TH, Thomas JD, Lauer MS. Chronotropic response to exercise predicts angiographic severity in patients with suspected or stable coronary artery disease. Am J Cardiol 1995;76:1228–32. [19] Lachterman B, Lehmann KG, Detrano R, Neutel J, Froelicher VF. Comparison of ST segment / heart rate index to standard ST criteria for analysis of exercise electrocardiogram. Circulation 1990;82(1):44–50. [20] Rozanski A, Diamond GA, Forrester JS, Berman DS, Morris D, Swan HJC. Alternative referent standards for cardiac normality. Ann Intern Med 1984;101:164–71. [21] Okin PM, Kligfield P. Population selection and performance of the exercise ECG for the identification of coronary artery disease. Am Heart J 1994;127(2):296–304. [22] Cannon RO, Epstein SE. ‘Microvascular angina’ as a cause of chest pain with angiographically normal coronary arteries. Am J Cardiol 1988;61(15):1338–43. [23] Philbrick JT, Horwitz RI, Feinstein AR. Methodologic problems of exercise testing for coronary artery disease: Groups, analysis and bias. Am J Cardiol 1980;46:807–13. [24] Diamond GA, Forrester JS. Analysis of probability as an aid in the clinical diagnosis of coronary-artery disease. N Engl J Med 1979;300(24):1350–8.
