Comparison between manual and computer measurement of ST-segment amplitude during exercise

Comparison between manual and computer measurement of ST-segment amplitude during exercise

JOURNAL OF ELECTROCARDIOLOGY, Supplemental Issue 1988, S130-S133 Comparison Between Manual and Computer Measurement of ST-Segment Amplitude During E...

441KB Sizes 0 Downloads 33 Views

JOURNAL OF ELECTROCARDIOLOGY, Supplemental

Issue 1988, S130-S133

Comparison Between Manual and Computer Measurement of ST-Segment Amplitude During Exercise BYROGERM. BOYLE,YERWA NIVEDITHA, CHRISTINA WINTER,ANDDAVIDA. S. G. MARY

SUMMARY Several studies have shown that the maximal ST/HR slope may be used as a reliable index of myocardial ischemia as assessed by coronary angiography, but this involves laborious training and derivation, particularly with respect to the measurement of ST segment amplitude, which is obtained by averaging values measured in at least 10 cardiac cycles in the steady state. The authors compared manual measurement of ST-segment amplitude with computer-processed beat using cardiac cycles in six consecutive patients with standard la-lead records obtained over 5 seconds and a beat processed by the recorder to represent each lead (modal beat, over 10 seconds). All recordings were made in the steady state. Two patients had myocardial ischemia, as assessed by means including the maximal ST/HR slope and the occurrence of ST-segment depression at the end of exercise. Comparisons were made between measurements in 324 pairs of STsegment amplitude obtained, respectively, from manually averaged recorded beats (average beat) and the modal beat during each step of the exercise test. The level of the ST-segment, (80 msec after the end of QRS complex) was independently obtained from the two records in a blinded fashion. The group data showed that the modal beat gave significantly lcwer values of ST-segment amplitude than the average beat. Similar results were obtained when 286 pairs of positive amplitudes were compared in the range of O-7.25 mm. In the remaining 38 pairs in which ST-segment depression was found, the amplitude in the modal beats was not significantly different from the average beats. When the amplitude was assessed in lo-12 average beats of cardiac cycles in 45 pairs, the protocol for obtaining the maximal ST/HR slope, no significant difference was found between average beats of 0.59 mm and modal beats of 0.69 mm. Cardiac cycles less than 10 were assessed to examine the influence of ST-segment amplitude and small numbers of recorded beats. Irrespective of the number of beats, amplitudes of values less than 2.45 mm in average beats were not significantly different from those in modal beats. The results show that the ECG recorder processing accuracy was better in cardiac beats with ST-segments having a low amplitude or a negative value than a highly positive amplitude. This should be useful in assessments of the level of the ST-segment at the end of exercise.

The maximal ST/HR slope has been shown to be an index of myocardial ischemia during exercise testing. l-3 In patients with angina or in From York District Hospital, York, England. Reprint requests to: Roger M. Boyle, York District Hospital, Wigginton Road, York Y03 7HE, England.

asymptomatic factory populations, the ST/HR relationship was shown to be a more accurate index than other exercise criteria.le6 However, determining the maximal ST/HR slope requires laborious measurement of the amplitude of the STsegment and training to achieve adequate recordings.l*’ s130

s131

ST-SEGMENT MEASUREMENT This was a preliminary study to explore the feasibility of using available ECG recorders incorporating computer software to assist in measuring ST-segment amplitude.

MATERIALS AND METHODS Six patients, five men and one woman, aged 49-70 years, who were randomly referred to the exercise laboratory in one outpatient session were examined. Two patients had myocardial ischemia as assessed by means including the maximal ST/HR slope and the occurrence of ST-segment depression of greater than 0.1 mV at the end of exercise. The exercise tests were supervised by one observer, according to a previously established protocol for obtaining maximal ST/HR slope.lr’ A commercially available ECG recorder (Siemens Elema, Sweden, Mingograf-740) was used. During the exercise test, recordings of the standard 12 leads were obtained during steps of steady-state periods, as previously described.lz’ At each step, the recordings comprised (1) &second recordings of the six limb leads and six chest leads (labelled “recorded beats”) and (2) single beats for each of the 12 leads, processed by the software of the recorder over the lo-second period during which the other beats were recorded (labelled “modal beats”). The amplitude of the ST-segment was measured by two observers in all cardiac beats, without knowledge of the identity of the recordings. The amplitude of the ST-segment (80 msec after the end of QRS complex) was measured independently in a blinded fashion in the recorded and modal beats. The measurements were as previously described.‘,’ Therefore, two values were available for analysis. For each step of the exercise test, the amplitude of the ST-segment in the recorded beats was averaged (average beat) and compared with that for the modal beat. There was nothing to indicate that obtaining the two beats during different time periods caused significant differences in ST-segment amplitude. All recordings were made in the steady state. The variability of measuring ST-segment amplitude at different times in the recorded beats was within the variability of measuring this amplitude. In terms of 95% tolerance limits of differences between measurements, the latter amounted to 0.15 mm. The maximal ST/HR slope was not used in the analysis, since it could not be obtained from the standard 12 ECG leads.‘*’ The investigation was designed to assess ST-segment amplitude of the modal beat by comparing it with that of the average beat. The latter has previously been used to derive the maximal ST/HR slope, particularly by obtaining the measurement in 10 cardiac cycles or more.“’ Student’s t-test for paired data was used for comparison. Group data are presented as mean * SEM.

RESULTS There were 324 pairs of ST-segment amplitude obtained from the average beat and the modal beat, as assessed during all steps of the exercise test in the six patients. Of these, the ST-segment amplitude value was negative in 38 and positive in 286. In addition, of the 324 pairs, 45 average beat values were obtained from lo-12 cardiac cycles; the remaining 279 values belonged to average beats representing 3-9 cardiac cycles. The individual values of all of the pairs are shown in Figure 1. The modal beat gave significantly lower (p < 0.0005) values of ST-segment amplitude of 0.55 mm t 0.039 (range, - 1.5-2.9) than the average beat values of 0.84 mm + 0.064 (range, - 1.7-7.25). This group underestimation by the modal beat amounted to 0.29 mm t 0.05. In a comparison of ST-segments with positive amplitude, the modal beat values of 0.65 mm k 0.039 (range, O-2.9) was significantly lower (p < 0.0005) than that of the average beat of 0.99 mm + 0.067 (range, 0- 7.25). This underestimation amounted to 0.34 mm I~I0.55. In the presence of negative ST-segment values, the values of the modal beat of - 0.23 mm t 0.070 (range, - 1.5 - 0.9) were not significantly different (p > 0.20) from those of the average beat of -0.29 + 0.075 (range, - 1.7-0.5). There was agreement between the group data of the two measurements. There were differences between 8

7 ST

segment

amplitude

I

b

e ?

---

s 8

mm Ruerage

beat

Fig. 1. Values of ST-segment amplitude of the modal beat versus those of the average beat. The continuous line is the line of no differences,

S132

BOYLE ET AL

individual pairs that were not systematic; the 95% tolerance limits of these differences amounted to 1.3 mm. There were no significant differences between modal and average beat values obtained from lo12 cardiac cycles. Collectively, the modal beats values were 0.69 mm k 0.17 (range, -0.7-5.4) and were not significantly different (p > 0.10) from those of the average beat of 0.59 mm or 0.15 (range, - 1.7-2.9). This group agreement between the two measurements contained individual differences. Of the 45 pairs, 44 values were less than 3 mm. The remaining value was 5.4 mm. Considering all 45 pairs, the 95% tolerance limits of individual differences between pairs amounted to 1.3 mm. To assess any effect of greatly positive ST-segment amplitude, a series of comparisons including numbers of cardiac cycles of the average beat of nine cycles (43 pairs), eight cycles (100 pairs), seven cycles (75 pairs), and six cycles (42 pairs) were made (Table I). The remaining 29 pairs, which represented three, four, and five cardiac cycles were considered not to have a number of pairs or cardiac cycles adequate for meaningful analysis. In the face of a possible decrease in precision using a small number of cardiac cycles, the recorder’s modal beat values were similar to those of the average beat only when the values were less than 2.5 mm. The majority of values of ST-segment amplitude of the average beat were less than 3 mm (Fig. 1). Amplitudes greater than 5 mm were obtained exclusively in chest leads VI-V3 in two patients. The recorded beats showed an early and a steep rise of the T wave. The recorder’s modal beat amplitude of these was systematically lower, and this underestimation was associated with decreases in the amplitude of the T wave.

DISCUSSION The accuracy of the maximal STlHR slope as an index of myocardial ischemia has depended on the relation between indicators of myocardial oxygen consumption and ischemia.1-3,6-g The heart rate has a large effect on myocardial consumption, and over decades manual measurement of the ST-segment has been used to assess myocardial ischemia. This accuracy has been reflected by minimal variability attending the derivation of the slope.lz2 In this derivation, the variability attending measurement of ST-segment amplitude is a major and crucial factor, since the heart rate can be measured with accuracy and the mathematical derivations involve averaging pr0cesses.l These considerations have led to meticulous care in obtaining adequate ECG recordings and measurement of ST-segment amplitude, which in turn have required laborious training and prolonged measurement times. 1,2Indeed, attempts to expedite measurement of ST-segment amplitude have led to reductions in the accuracy of the maximal ST/HR slope. This has included the use of a locally developed microcomputer interfaced with an adapted ECG rec0rder.l’ Computers have been used successfully to assist in the analysis of the STiHR slope.ll This study was done to explore the use of an ECG recorder’s software to expedite measurement of ST-segment amplitude. Other ECG recorders are being examined at present and compared in larger, ongoing investigation. These findings show that the recorder’s software could reduce the time required to measure ST-segment amplitude. A better representation of amplitudes from the often-used average beat by the recorder’s assessment was found when the former did not have greatly positive amplitude or

TABLE I ST-Segment Amplitude Values Obtained From Less Than 10 Cardiac Cycles for Modal and Average Beat ST-Segment Amplitude (mm) No. of Cardiac Cycles

Modal

Average

Modal

9 8 7 6

0.79 0.48 0.67 0.32

1.22 0.87 0.95 0.38

0.12 0.05 0.08 0.12

SEM

Mean

Range Average 0.19

0.12 0.12 0.12

Modal

Average

P

0.0-2.7 - 0.6-2.4 - 0.4-2.9 - 1.5-2.3

0.08-5.3 - 0.21-7.3 - 1.21-4.3 - 1.27-2.5

O.lO

ST-SEGMENT MEASUREMENT ST-segment elevation. Note that the maximal ST/HR slope could be derived from positive ST-segment amplitudes, which during exercise progress linearly into a negative amplitude.lp’ Also, we have found the relation between ST-segment elevation and the heart rate during exercise to be useful in patients soon after myocardial infarction.12 The mechanism by which the recorder’s software systematically underestimated ST-segment elevations was not examined in this study. The data from patients with marked ST-segment elevation suggest that the recorder’s processing has led to changes in wave amplitude. In our experience in conventional ECG recorders, the use of filters to narrow the range of the frequency response and thus change waveforms alters the reference points used to define the site of assessing ST-segment amplitude. That the ECG recorder processing accuracy was better in cardiac beats with ST-segments with a low amplitude or a negative value than with those with a highly positive amplitude suggests that it should be useful in assessments of the level of the ST-segment at the end of exercise.

s133

excessive

REFERENCES LINDEN RJ, MARYDASG: Limitations and reliability of exercise electrocardiography tests in coronary heart disease. Cardiovasc Res 16:675, 1982 BISHOPN, ADLAKHA HL, BOYLERM ET AL: ST segment/heart rate relationship. an index of myocardial ischaemia. Int J on the

4.

5.

6.

7.

8.

9.

of the maximal in asymptomatic factory J Electrocardiol 20 (suppl): I, HAJDUCZKI E: Quantitative evaluation exercise-induced ST-segment depression for estimation of degree of coronary artery disease. Eur Heart J I, BERENYI I, ENGHOFF ETAL: Qualitative and quantitative evaluation of the exercise electrocardiogram in assessing the degree of of coronary heart disease. J Electrocardiol 1855, 1985 KLIGFIELD P, OKINPM, AMEISEN 0, BORERJS: Evaluation of coronary artery disease by an improved method of exercise electrocardiography: the ST segment/heart rate slope. Am Heart J 112:589, 1986 FINKELHOR RS, NEWHOUSE KE, VROBELTRETAL:The ST segment/heart rate slope as a predictor of coronary artery disease: comparison with quantitative thallium imaging and conventional ST segment criteria. Am Heart J 112:296, 1986 DETRANO R, SALCEDO E, PASSALACQUA M, FRIISR: Exercise electrocardiographic variables: a critical appraisal, J Am Co11Cardiol 8:836, 1986 CRANERA, LINDENRJ, MARYDASG: A computer to analyse the relationship between ST segment depression and heart rate during exercise. J Physiol 330:15P, 1982 OKINPM, AMEISEN 0, KLIGFIELD P: A modified treadmill exercise protocol for computer-assisted analysis of the ST segment/heart rate slope: methods and reproducibility. J Electrocardiol 19:311, 1986 BISHOP N, MACKINTOSH AF, STOKER JB ETAL: Exercise-induced ST segment elevation related to heart rate: correlation with left ventricular (LV) and coronary angiography following recent myocardial infarction (MI). Clin Sci 73 (suppl 17):39P, 1987