- Email: [email protected]
heart rate analysis of the exercise electrocardiographic test in asymptomatic middle-aged population
so
60
70
-
moupl
-
moup2
60
60
natients at risk should be Dated at 290 beats/min when they leave the elect&physiology laboratory, at least for 24 hours. Whether these patients remain at higher risk of sudden death is unclear, although none had an arrhythmia before ablation, and in 13 + 3 months of follow up, none have had a recurrent life-threatening ventricular arrhythmia. Our data suggest that increased QT dispersion is a marker of electrical instability and predicts PMVTM? after AVNA, particularly in association with depressed left ventricular systolic function. Patients showing marked QT dispersion require careful monitoring after AVNA. 100
Paced Rate (bpm)
sudden cardiac death to those who did not suggest that patients with increased QT dispersion and depressed left ventricular function may be at substantially increased risk for ventricular arrhythmias after ablation. We therefore recommend that
1. JackmanWM. Wang X, Friday RJ, Roman CA, Moulton KP, Beckman KJ, McClelland JH, Twidale N, Hazlett HA, Prior MJ et al. Catheter ablation of accessory atrioventricular pathway (Wolff-Parkinson-White Syndrome) by radiofrequency current. N Engl J Med 1991;324:1605-1611. 2. PetersRHJ, Wever EFD, Hatter RNW, Wittkampf FHM, Robles de Medina EOR. Bradycardia dependentQT prolongation and ventricular fibrillation following catheter ablation of the atrioventricular junction with radiofrequency energy. PACE 1994;17:108-112. 3. Statters DJ, Malik M, Ward DE, Camm JA. QT dispersion: problems of methodology and clinical significance. .I Cardiovasc Electrophysiol 1994,s: 672-685. 4. Day CP, McComb JM, Campbell RWF. QT dispersion in sinus beats and ventricular extrasystoles in normal hearts. Br Heart J 1992;67:39-
41. 5. priori SG, Napolitano C, Diehl L, Schwartz PJ.Dispersion of the QT interval. A marker of therapeutic efficacy in the idiopathic long QT syndrome. Circulation 1994;89:1681-1689.
Reproducibtlity of the ST-Segment Depreosion/Heart raphic Test Rate Analysis of the Exercise Electrocardio Popu r ation in Asymptomatic Middle-Aged Rami Lehtinen, LicTech, Harri Sievenen, PhD, Jari Viik, LicTech, llkka Vuori, MD, and Jaakko Malmivuo, PhD
t has been shown that noninvasive diagnosis of coronary artery diseasecan be improved by anaIlyzing the ST-segment depression against heart rate (ST/HR) during the exercise electrocardiographic (ECG) test.‘.’ Recently, the STIHR hysteresis,a variable that integrates the diagnostic information of both the exerciseand recovery phasesof the test, was found to be the most accurate diagnostic variable based on the STA-IR analysiss3However, further refinement of the ST/HR analysis would require adeFrom the Ragnar Granit Institute, Tampere University of Technolo y’ and the UKK Institute for Health Promotion Research, Tampere, ?inl land. This work has been financially supported b y..the Academy of Finland, Helsinki; Alfred Kordelin Foundation, Hesmkr; Finmsh Cultural Foundation [Fund of Pirkanmaa), Tampere; Finnish Cultural Foundotion, Finnish Foundation for Cardiovascular Research, Helsinki; Ragnar Granit Foundation, Helsinki; Tampere Science Foundation, Tampere; and Wihuri Foundation, Helsinki, Finland. Dr. tehtinen’s address is: Ragnar Granit Institute, Tampere University of Technology, PG. Box 692, FIN-33 10 1 Tampere, Finland. Manuscript received September 16, 1996; revised manuscript received and accepted January 27, 1997. 1414
01997 by Excerpta Medica, All rights reserved.
Inc.
quate knowledge of inherent (e.g., method-induced) nondiagnostic variability in the above variables.This information on reproducibility has been scarceso far and limited to patients with coronary artery disease.4*5Knowing especially the magnitude of intraindividual variability in healthy asymptomatic populations would be crucial if one wishes to determine whether a longitudinal change in a diagnostic variable is large enough to indicate that a change with real diagnostic value has happened. Therefore, the objective of this study was to determine the reproducibility of the STLHR hysteresis, ST/HR index, and end-exerciseST depressionin an asymptomatic middle-aged population, an age-cohort that is most often referred to exercise ECG tests. ... Maximal exercise ECG tests were performed twice within a period of 6 to 8 months in 61 middleaged (5 1 to 54 years) healthy volunteers (28 men and 33 women). The subjects gave informed consent before the study. Each subject was sedentary (vigorous 0002-9149/97/Q 17.00 PII SOOO2-9149(97)00155-O
TABLE I Description of the Study Population Variables During the First Test Characteristic Age b) Height (cm] Weight [kg) Maximum heart rate achieved (beats/mm) Respiratory quotient ST/HR hysteresis (mV) ST/HR index (hV/beots/min) Endexercise ST depression (mV)
and’ Exercise
Men (n = 28) 52 r 1 177?6 81.6 + 8.7 172 + 12 1.17 -0.029 1.28 0.08
+ 2 + +
immediately before starting the exercise with the patient sitting on the bicycle, at the end of each minute of Women (n = 33) exercise, at the end-exercise, and every 12 seconds during the first 3 53 2 1 161 +4 consecutive minutes of postexercise 64.8 + 7.6 . recovery. A continuous piecewise 1732 10 linear function was obtained by connecting the consecutive ST/HR 1.12 t 0.05 -0.011 t 0.012 data pairs of the exercise phase 1.22 + 0.76 with lines. Similarly, a continuous 0.09 k 0.07 piecewise linear function for the postexercise recovery phase was constructed by connecting the consecutive ST/HR data pairs of the first 3 minutes of recovery phase starting from the ST/HR data pair at the end-exercise. Then, the difference between the above curves was integrated over the HR from the minimum HR of recovery to the maximum HR. Finally the ST/HR hysteresis was obtained by dividing the integrated net difference by the HR difference of the integration interval in order to normalize the result with respect to the postexercise HR decrement. The ST/HR index was calculated as the gradient between the ST/HR pairs at the start-exerciseand at the end-exercise as suggested by Detrano and coworkers.7 All continuous data were expressed as mean + SD. The reproducibility of the exercise ECG variables between the repeated measurementswas determined as recommended by Bland and Altman.* Definition of reproducibility was ? 1.96 times the SD of the differences between the pairs of measurement (SD,,) using the same method. This range corresponds to 95% limits of agreementwithin which intra-individual changes should be considered nonsignificant due to inherent variability of the method. As another measureof reproducibility, the agreementof interpretation between the repeated measurements was defined as the percentage of the subjects in which the interpretation of both measurementswas the same. It was determined separately for each of the 3 exercise ECG variables using specific fixed partition values that provided equal specificity of 80% for each method in detecting coronary artery disease in our previous study.3 These positive test criteria were ST/I-IR hysteresis rO.O1OmV, ST/HR index 2 1.60 pV/beats/min, and end-exerciseST depression rO.10 mV. The agreements of interpretations between the repeated measurementswith different variables were compared by McNemar modification of the chi-squared test. Because of the 3 comparisons, the p value <0.017 (Bonferroni correction) was required for rejection of null hypothesis. Maximal exercise tests were successful at both testing sessions(maximum HR achieved [ 172 2 11 vs 172 5 11 beats/min], RQ [1.15 + 0.06 vs 1.14 + 0.071). The reproducibilities were _+0.040 mV, t 1.24 pV/beats/min, and +O.11 mV for the ST/HR hysteresis, ST/HR index and end-exercise ST depression, respectively. The Bland-Altman plots are
Electrocardiographic
0.06 0.012 0.81 0.08
Continuous data ore mean z SD.
exercise no more than twice a week), nondieting, nonsmoking, and not excessively obese (body mass index <33). One month before the first maximal test, the subjects accustomed themselves to the testing procedure by performing a submaximal test. According to careful medical screening done before both maximal tests, none of these patients had an abnormal electrocardiogram at rest, or had a history or symptoms of cardiovascular, musculoskeletal, respiratory, or other chronic diseasesthat might limit the maximal exercise testing. Between the repeated teststhe subjectswere asked to keep their living habits unchanged, and during the study period none of the subjects showed any clinical signs of evolving heart or other disease.The clinical data of the study group are given in Table I. All subjects were tested on a bicycle ergometer in the exercise physiology laboratory of the UKK Institute (Tampere, Finland). Each test was conducted by 1 of the 3 experienced technicians of the institute. The initial workload was 10 W for women and 20 W for men, and was increased by 10 and 20 W every minute for women and men, respectively. The ECG recordings and HR measurementswere done with a Marquette CASE 12 recorder (Marquette Inc., Milwaukee, Wisconsin). The respiratory quotient (RQ) was used as an index of maximality of the exercise ECG test, and was determined as the ratio between the carbon dioxide produced and oxygen consumed measuredby an automatic ventilatory gas analyzer (Sensomredics 29002, Sensormedics Inc., Anaheim, California). An RQ value above 1.00 indicates anaerobic energy production. With the computerized analysesof the stored ST-segmentand HR data,(jthe maximum values of the ST/HR hysteresis, ST/HR index, and end-exercise ST depression were determined from the Mason-Likar modification of the 12-lead system (leads aVL, aVR, and Vi excluded) for each subject. The ST-segmentdepression required for theseanalyseswas measuredto the nearest 10 PV at 60 ms after the QRS-offset. Of note, an ST-segment elevation was expressed as a negative ST-segment depression. The ST/I-IR hysteresis for each lead was determined as described recently.3 In short, the ST/HR diagram containing the pairs of the ST depression and HR was determined by measuring the given data
BRIEFREPORTS 1415
interpretation between the repeatedmeasurementsof the ST/HR hysteresis was significantly better than that of the ST/HR index (p = 0.0045) or the endexercise ST depression (p = O.OOlO),whereas no significant difference was observedbetween the ST/ HR index and the end-exercise ST depression (p = 0.32). . . .
“.“”
-0.05 -0.04 -0.03 -0.02 -0101 0.00 0.01 0.02
of ST/HR hysteresis (mV) 0 .--....__.
. .
---
.--- - .._---
sy--
+l.% SD,, -.-..-i~07
. -0
0 0 0 0.. 8, 0 0 0 u*-
n
0%
0 080’ 99
0
.O O.0 l
.*
Mp -0.17
a.O 0
’
1 :E]-+=yi 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.m
Meanvalueof ST/HR index (pV@pm) 0.15 0.10
ss‘ .iJ
0.05
1 E
0.00
Acknowledgment: We thank Tiina Mets&an@ MA, for technical assistancein processing the data.
.g -0.05 8 8 g Q
-0.10 -0.15
-l%SD -0.12
..---.--.---o------.-L._.0
.
0 J -0.20 -0.05 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
Mean value of ST,, (mV) FIGURE 1. Bland-Altman reproducibility plok of the ST/HR h skresis, ST/HR,index, and end-exercise ST depression (ST4 le. tween repeated exercise ECG tests. CkKd and open circles, the data points of men and women, respectively. The differences in the exercise ECG voriaMes between the repeated measurements (vertical axes) was cakulclted by subtracting the value of the second test from that of the first test. bpm = beats per minute; SD, = SD of the dihnce between the repeated measurements.
presented in Figure 1. The agreement of interpretation between the repeated measurementswas 97% (59 of 61), 79% (48 of 61), and 74% (45 of 61) for the ST/HR hysteresis, ST/HR index, and end-exercise ST depression, respectively. The agreement of
1416
Given the large intra-individual variability of ECG recordings at rest:” the reproducibility of the exercise ECG variables is apparently bound to be relatively poor, a fact that was actually confirmed in this study in asymptomatic middle-aged individuals. The magnitude of change in the variable, which has to be observed to make the clinician confident that a real diagnostic change has occurred, was surprisingly large as indicated by the Bland-Altman plots in Figure 1. However, the possibility that some of the subjectsmay have developeda significant change in their coronary statusis not definitely ruled out, but given that coronary atherosclerosisoccurs over a period of decades, one would not expect to find any substantial coronary changes in 6 to 8 months. It is worth noting that the agreement of interpretation with the ST/HR hysteresis appeared to be significantly better than those with the ST/HR index or endexercise ST depression. The findings of this study on reproducibility of the exercise ECG variables in 61 asymptomatic middle-aged subjects give further support to the clinical utility of the ST/HR hysteresis. The results also indicated that the observed change in the exercise ECG variable between repeated measurements must be large to make the clinician confident that a real diagnostic change has occurred.
THE AMERICAN JOURNAL OF CARDIOLOGY@ VOL. 79
1. Okin PM, Kligfield P. Heart rate adjustmentof ST segmentdepressionand performance of the exercise electrocardiogram: a critical evaluation. J Am Coil Cardiol 1995;25:1726-1735. 2. Lehtinen R, Sievshen H, Uusitalo A, Niemek K, Tmjamnaa V, Malmivuo J. Performance characteristics of various exercise electrocardiographic claasifiers in different clinical populations. J Elecrrocardiol 1994;27:1l-22. 3. Lehtinen R, Sievahen H, Viii J, Tmjanmaa V, Niemell K, Mahnivuo J. Accurate detection of coronary artery diseaseby integrated analysis of the STsegmentdepression/heartrate patterns during the exercise and recovery phases of the exercise electrocardiography test. Am J Cardiol 1996;78:1002-1006. 4. Eltin M, Mary D, Smith D, Linden R. Prediction of severity of coronary heart disease using slope of submaximal ST-segment/heartrate relationship. Cardiovasc Res1980;14:681-691. 5. Okin PM, Ameisen 0, Kligfield P. A modified treadmill exercise protocol for computer assistedanalysis of the ST segment/heartrate slope: methodsand reproducibiiity. J Electrocardiol 1986;19:31l-318. 6. Lehtinen R, Xnttinen H, Sievshen H, Mahnivuo J. A computerprogram for comprehensiveST-segmentdepression/heartrate analysis of the exercise ECG test. CompurMeth Prog Bio 1996;50:63-71. 7. Detrano R, Sakedo E, PassalacquaM, Friis R. Exercise electrocardiographic variables: a critical appraisal. JAm CON Cardiol 1986;8:836-847. 8. Bland JM, Altman DG: Statistical methodsfor assessingagreementbetween two methodsof clinical measurement.Lancer 1986$X:307-310. 9. Michaels L, Cadoret RJ. Day-to-day variability in the normal electrocardiogram. Er Hean J 1%7;29:913-919. 10. Willems JL, Poblete PF, Pipberger HV. Day-to-day-variation of the normal orthogonal electrxardiogram and vectorcardiogram. Circukzh’on 1972; 451057-1064.
MAY 15, 1997
60
70
-
moupl
-
moup2
60
60
natients at risk should be Dated at 290 beats/min when they leave the elect&physiology laboratory, at least for 24 hours. Whether these patients remain at higher risk of sudden death is unclear, although none had an arrhythmia before ablation, and in 13 + 3 months of follow up, none have had a recurrent life-threatening ventricular arrhythmia. Our data suggest that increased QT dispersion is a marker of electrical instability and predicts PMVTM? after AVNA, particularly in association with depressed left ventricular systolic function. Patients showing marked QT dispersion require careful monitoring after AVNA. 100
Paced Rate (bpm)
sudden cardiac death to those who did not suggest that patients with increased QT dispersion and depressed left ventricular function may be at substantially increased risk for ventricular arrhythmias after ablation. We therefore recommend that
1. JackmanWM. Wang X, Friday RJ, Roman CA, Moulton KP, Beckman KJ, McClelland JH, Twidale N, Hazlett HA, Prior MJ et al. Catheter ablation of accessory atrioventricular pathway (Wolff-Parkinson-White Syndrome) by radiofrequency current. N Engl J Med 1991;324:1605-1611. 2. PetersRHJ, Wever EFD, Hatter RNW, Wittkampf FHM, Robles de Medina EOR. Bradycardia dependentQT prolongation and ventricular fibrillation following catheter ablation of the atrioventricular junction with radiofrequency energy. PACE 1994;17:108-112. 3. Statters DJ, Malik M, Ward DE, Camm JA. QT dispersion: problems of methodology and clinical significance. .I Cardiovasc Electrophysiol 1994,s: 672-685. 4. Day CP, McComb JM, Campbell RWF. QT dispersion in sinus beats and ventricular extrasystoles in normal hearts. Br Heart J 1992;67:39-
41. 5. priori SG, Napolitano C, Diehl L, Schwartz PJ.Dispersion of the QT interval. A marker of therapeutic efficacy in the idiopathic long QT syndrome. Circulation 1994;89:1681-1689.
Reproducibtlity of the ST-Segment Depreosion/Heart raphic Test Rate Analysis of the Exercise Electrocardio Popu r ation in Asymptomatic Middle-Aged Rami Lehtinen, LicTech, Harri Sievenen, PhD, Jari Viik, LicTech, llkka Vuori, MD, and Jaakko Malmivuo, PhD
t has been shown that noninvasive diagnosis of coronary artery diseasecan be improved by anaIlyzing the ST-segment depression against heart rate (ST/HR) during the exercise electrocardiographic (ECG) test.‘.’ Recently, the STIHR hysteresis,a variable that integrates the diagnostic information of both the exerciseand recovery phasesof the test, was found to be the most accurate diagnostic variable based on the STA-IR analysiss3However, further refinement of the ST/HR analysis would require adeFrom the Ragnar Granit Institute, Tampere University of Technolo y’ and the UKK Institute for Health Promotion Research, Tampere, ?inl land. This work has been financially supported b y..the Academy of Finland, Helsinki; Alfred Kordelin Foundation, Hesmkr; Finmsh Cultural Foundation [Fund of Pirkanmaa), Tampere; Finnish Cultural Foundotion, Finnish Foundation for Cardiovascular Research, Helsinki; Ragnar Granit Foundation, Helsinki; Tampere Science Foundation, Tampere; and Wihuri Foundation, Helsinki, Finland. Dr. tehtinen’s address is: Ragnar Granit Institute, Tampere University of Technology, PG. Box 692, FIN-33 10 1 Tampere, Finland. Manuscript received September 16, 1996; revised manuscript received and accepted January 27, 1997. 1414
01997 by Excerpta Medica, All rights reserved.
Inc.
quate knowledge of inherent (e.g., method-induced) nondiagnostic variability in the above variables.This information on reproducibility has been scarceso far and limited to patients with coronary artery disease.4*5Knowing especially the magnitude of intraindividual variability in healthy asymptomatic populations would be crucial if one wishes to determine whether a longitudinal change in a diagnostic variable is large enough to indicate that a change with real diagnostic value has happened. Therefore, the objective of this study was to determine the reproducibility of the STLHR hysteresis, ST/HR index, and end-exerciseST depressionin an asymptomatic middle-aged population, an age-cohort that is most often referred to exercise ECG tests. ... Maximal exercise ECG tests were performed twice within a period of 6 to 8 months in 61 middleaged (5 1 to 54 years) healthy volunteers (28 men and 33 women). The subjects gave informed consent before the study. Each subject was sedentary (vigorous 0002-9149/97/Q 17.00 PII SOOO2-9149(97)00155-O
TABLE I Description of the Study Population Variables During the First Test Characteristic Age b) Height (cm] Weight [kg) Maximum heart rate achieved (beats/mm) Respiratory quotient ST/HR hysteresis (mV) ST/HR index (hV/beots/min) Endexercise ST depression (mV)
and’ Exercise
Men (n = 28) 52 r 1 177?6 81.6 + 8.7 172 + 12 1.17 -0.029 1.28 0.08
+ 2 + +
immediately before starting the exercise with the patient sitting on the bicycle, at the end of each minute of Women (n = 33) exercise, at the end-exercise, and every 12 seconds during the first 3 53 2 1 161 +4 consecutive minutes of postexercise 64.8 + 7.6 . recovery. A continuous piecewise 1732 10 linear function was obtained by connecting the consecutive ST/HR 1.12 t 0.05 -0.011 t 0.012 data pairs of the exercise phase 1.22 + 0.76 with lines. Similarly, a continuous 0.09 k 0.07 piecewise linear function for the postexercise recovery phase was constructed by connecting the consecutive ST/HR data pairs of the first 3 minutes of recovery phase starting from the ST/HR data pair at the end-exercise. Then, the difference between the above curves was integrated over the HR from the minimum HR of recovery to the maximum HR. Finally the ST/HR hysteresis was obtained by dividing the integrated net difference by the HR difference of the integration interval in order to normalize the result with respect to the postexercise HR decrement. The ST/HR index was calculated as the gradient between the ST/HR pairs at the start-exerciseand at the end-exercise as suggested by Detrano and coworkers.7 All continuous data were expressed as mean + SD. The reproducibility of the exercise ECG variables between the repeated measurementswas determined as recommended by Bland and Altman.* Definition of reproducibility was ? 1.96 times the SD of the differences between the pairs of measurement (SD,,) using the same method. This range corresponds to 95% limits of agreementwithin which intra-individual changes should be considered nonsignificant due to inherent variability of the method. As another measureof reproducibility, the agreementof interpretation between the repeated measurements was defined as the percentage of the subjects in which the interpretation of both measurementswas the same. It was determined separately for each of the 3 exercise ECG variables using specific fixed partition values that provided equal specificity of 80% for each method in detecting coronary artery disease in our previous study.3 These positive test criteria were ST/I-IR hysteresis rO.O1OmV, ST/HR index 2 1.60 pV/beats/min, and end-exerciseST depression rO.10 mV. The agreements of interpretations between the repeated measurementswith different variables were compared by McNemar modification of the chi-squared test. Because of the 3 comparisons, the p value <0.017 (Bonferroni correction) was required for rejection of null hypothesis. Maximal exercise tests were successful at both testing sessions(maximum HR achieved [ 172 2 11 vs 172 5 11 beats/min], RQ [1.15 + 0.06 vs 1.14 + 0.071). The reproducibilities were _+0.040 mV, t 1.24 pV/beats/min, and +O.11 mV for the ST/HR hysteresis, ST/HR index and end-exercise ST depression, respectively. The Bland-Altman plots are
Electrocardiographic
0.06 0.012 0.81 0.08
Continuous data ore mean z SD.
exercise no more than twice a week), nondieting, nonsmoking, and not excessively obese (body mass index <33). One month before the first maximal test, the subjects accustomed themselves to the testing procedure by performing a submaximal test. According to careful medical screening done before both maximal tests, none of these patients had an abnormal electrocardiogram at rest, or had a history or symptoms of cardiovascular, musculoskeletal, respiratory, or other chronic diseasesthat might limit the maximal exercise testing. Between the repeated teststhe subjectswere asked to keep their living habits unchanged, and during the study period none of the subjects showed any clinical signs of evolving heart or other disease.The clinical data of the study group are given in Table I. All subjects were tested on a bicycle ergometer in the exercise physiology laboratory of the UKK Institute (Tampere, Finland). Each test was conducted by 1 of the 3 experienced technicians of the institute. The initial workload was 10 W for women and 20 W for men, and was increased by 10 and 20 W every minute for women and men, respectively. The ECG recordings and HR measurementswere done with a Marquette CASE 12 recorder (Marquette Inc., Milwaukee, Wisconsin). The respiratory quotient (RQ) was used as an index of maximality of the exercise ECG test, and was determined as the ratio between the carbon dioxide produced and oxygen consumed measuredby an automatic ventilatory gas analyzer (Sensomredics 29002, Sensormedics Inc., Anaheim, California). An RQ value above 1.00 indicates anaerobic energy production. With the computerized analysesof the stored ST-segmentand HR data,(jthe maximum values of the ST/HR hysteresis, ST/HR index, and end-exercise ST depression were determined from the Mason-Likar modification of the 12-lead system (leads aVL, aVR, and Vi excluded) for each subject. The ST-segmentdepression required for theseanalyseswas measuredto the nearest 10 PV at 60 ms after the QRS-offset. Of note, an ST-segment elevation was expressed as a negative ST-segment depression. The ST/I-IR hysteresis for each lead was determined as described recently.3 In short, the ST/HR diagram containing the pairs of the ST depression and HR was determined by measuring the given data
BRIEFREPORTS 1415
interpretation between the repeatedmeasurementsof the ST/HR hysteresis was significantly better than that of the ST/HR index (p = 0.0045) or the endexercise ST depression (p = O.OOlO),whereas no significant difference was observedbetween the ST/ HR index and the end-exercise ST depression (p = 0.32). . . .
“.“”
-0.05 -0.04 -0.03 -0.02 -0101 0.00 0.01 0.02
of ST/HR hysteresis (mV) 0 .--....__.
. .
---
.--- - .._---
sy--
+l.% SD,, -.-..-i~07
. -0
0 0 0 0.. 8, 0 0 0 u*-
n
0%
0 080’ 99
0
.O O.0 l
.*
Mp -0.17
a.O 0
’
1 :E]-+=yi 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.m
Meanvalueof ST/HR index (pV@pm) 0.15 0.10
ss‘ .iJ
0.05
1 E
0.00
Acknowledgment: We thank Tiina Mets&an@ MA, for technical assistancein processing the data.
.g -0.05 8 8 g Q
-0.10 -0.15
-l%SD -0.12
..---.--.---o------.-L._.0
.
0 J -0.20 -0.05 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
Mean value of ST,, (mV) FIGURE 1. Bland-Altman reproducibility plok of the ST/HR h skresis, ST/HR,index, and end-exercise ST depression (ST4 le. tween repeated exercise ECG tests. CkKd and open circles, the data points of men and women, respectively. The differences in the exercise ECG voriaMes between the repeated measurements (vertical axes) was cakulclted by subtracting the value of the second test from that of the first test. bpm = beats per minute; SD, = SD of the dihnce between the repeated measurements.
presented in Figure 1. The agreement of interpretation between the repeated measurementswas 97% (59 of 61), 79% (48 of 61), and 74% (45 of 61) for the ST/HR hysteresis, ST/HR index, and end-exercise ST depression, respectively. The agreement of
1416
Given the large intra-individual variability of ECG recordings at rest:” the reproducibility of the exercise ECG variables is apparently bound to be relatively poor, a fact that was actually confirmed in this study in asymptomatic middle-aged individuals. The magnitude of change in the variable, which has to be observed to make the clinician confident that a real diagnostic change has occurred, was surprisingly large as indicated by the Bland-Altman plots in Figure 1. However, the possibility that some of the subjectsmay have developeda significant change in their coronary statusis not definitely ruled out, but given that coronary atherosclerosisoccurs over a period of decades, one would not expect to find any substantial coronary changes in 6 to 8 months. It is worth noting that the agreement of interpretation with the ST/HR hysteresis appeared to be significantly better than those with the ST/HR index or endexercise ST depression. The findings of this study on reproducibility of the exercise ECG variables in 61 asymptomatic middle-aged subjects give further support to the clinical utility of the ST/HR hysteresis. The results also indicated that the observed change in the exercise ECG variable between repeated measurements must be large to make the clinician confident that a real diagnostic change has occurred.
THE AMERICAN JOURNAL OF CARDIOLOGY@ VOL. 79
1. Okin PM, Kligfield P. Heart rate adjustmentof ST segmentdepressionand performance of the exercise electrocardiogram: a critical evaluation. J Am Coil Cardiol 1995;25:1726-1735. 2. Lehtinen R, Sievshen H, Uusitalo A, Niemek K, Tmjamnaa V, Malmivuo J. Performance characteristics of various exercise electrocardiographic claasifiers in different clinical populations. J Elecrrocardiol 1994;27:1l-22. 3. Lehtinen R, Sievahen H, Viii J, Tmjanmaa V, Niemell K, Mahnivuo J. Accurate detection of coronary artery diseaseby integrated analysis of the STsegmentdepression/heartrate patterns during the exercise and recovery phases of the exercise electrocardiography test. Am J Cardiol 1996;78:1002-1006. 4. Eltin M, Mary D, Smith D, Linden R. Prediction of severity of coronary heart disease using slope of submaximal ST-segment/heartrate relationship. Cardiovasc Res1980;14:681-691. 5. Okin PM, Ameisen 0, Kligfield P. A modified treadmill exercise protocol for computer assistedanalysis of the ST segment/heartrate slope: methodsand reproducibiiity. J Electrocardiol 1986;19:31l-318. 6. Lehtinen R, Xnttinen H, Sievshen H, Mahnivuo J. A computerprogram for comprehensiveST-segmentdepression/heartrate analysis of the exercise ECG test. CompurMeth Prog Bio 1996;50:63-71. 7. Detrano R, Sakedo E, PassalacquaM, Friis R. Exercise electrocardiographic variables: a critical appraisal. JAm CON Cardiol 1986;8:836-847. 8. Bland JM, Altman DG: Statistical methodsfor assessingagreementbetween two methodsof clinical measurement.Lancer 1986$X:307-310. 9. Michaels L, Cadoret RJ. Day-to-day variability in the normal electrocardiogram. Er Hean J 1%7;29:913-919. 10. Willems JL, Poblete PF, Pipberger HV. Day-to-day-variation of the normal orthogonal electrxardiogram and vectorcardiogram. Circukzh’on 1972; 451057-1064.
MAY 15, 1997