American Journal of Critical Care. 2004;13: 228-234
Copyright © 2004 by the American Association of Critical-Care Nurses.
COMPARISON OF SIGNAL QUALITY BETWEEN EASI AND MASON-LIKAR 12-LEAD ELECTROCARDIOGRAMS DURING PHYSICAL ACTIVITY
By
Annika Welinder, MD,
Leif Sörnmo, PhD,
Dirk Q. Feild, MA,
Charles L. Feldman, DSc,
Jonas Pettersson, MD, PhD,
Galen S. Wagner, MD and
Olle Pahlm, MD, PhD.
From the Department of Clinical Physiology, University Hospital, Lund, Sweden (AW, JP, OP); Department of Electroscience, Lund Institute of Technology, Lund, Sweden (LS); Philips Medical Systems, Oxnard, Calif (DQF), Brigham and Women’s Hospital, Boston, Mass (CLF), and Duke University Medical Center, Durham, NC (GW).
 |
Abstract
|
|---|
• Background Myoelectric noise and baseline wander, artifacts that appear when patients move during electrocardiographic monitoring, can cause false alarms. This problem can be addressed by using a reduced lead set and placing electrodes on the anterior part of the torso only. The Mason-Likar modification of the standard 12-lead electrocardiogram and the EASI lead system are 2 alternative systems for lead placement.
• Objectives To test the hypothesis that the EASI lead system is less susceptible to artifacts than is the Mason-Likar modification of the standard 12-lead electrocardiogram.
• Methods Baseline wander and myoelectric noise amplitudes of EASI and Mason-Likar 12-lead electrocardiograms were compared. Twenty healthy volunteers participated. Both lead systems were recorded simultaneously for different types of physical activities. For each lead in each subject, baseline wander and myoelectric noise were measured for both systems, at rest and during each physical activity.
• Results The outcome for baseline wander was mixed. For myoelectric noise content, the EASI system performed better for the limb leads in the different physical activities. In the precordial leads, the differences were minimal or mixed. However, for supine-to-right turning, EASI performed worse than the Mason-Likar system.
• Conclusions The 2 systems have similar susceptibilities to baseline wander. The EASI system is, however, less susceptible to myoelectric noise than is the Mason-Likar system. EASI performed worse than Mason-Likar for turning supine to right, because only the EASI system uses an electrode in the right-midaxillary line.
Monitoring electrocardiograms (ECGs) is useful in many in-hospital clinical situations as well as in everyday activities (Holter monitoring). One problem with ECG monitoring is its susceptibility to artifacts due to such problems as poor electrode contact or disturbances from skeletal muscle. Excessive artifacts have made it impractical to use the standard 12-lead ECG with distal limb leads (standard-limb) when monitoring. The Mason-Likar modification (ML) of the standard 12-lead ECG is commonly used for ECG monitoring.1 ML uses all the conventional precordial electrode sites, but the limb electrodes are connected to sites on the anterior part of the torso instead of to distal limb sites. Repositioning the recording electrodes to only torso sites achieves the goal of reducing artifacts but requires the same number of electrodes (10) as does the standard-limb system. An alternative to ML is to use only 3 to 5 electrodes. This number is sufficient for rhythm detection but provides limited information on waveform morphology.
ECG signals from a few well-defined torso positions, in combination with mathematical transformations, however, may provide a "derived" 12-lead ECG, which is useful also for analysis of morphology. The EASI system,2–4 introduced by Dower et al5 in the 1980s, uses 4 electrodes on easily located positions on the torso plus a fifth, ground electrode. Previous studies compared waveforms6 and diagnostic capability3,7–9 of 12-lead ECGs derived from EASI leads versus standard (proximal or distal limb lead placement) 12-lead ECGs. These comparisons were done with particular reference to monitoring patients in the intensive care environment. Studies4,7 suggest that the EASI system is less susceptible to artifacts than is the standard 12-lead ECG. To test this hypothesis, we designed a study to systematically compare baseline wander and myoelectric noise amplitudes in controlled situations between simultaneously recorded EASI-derived 12-lead ECGs and ML 12-lead ECGs.
|
Methods
|
|---|
Study Population
Twenty ostensibly healthy volunteers participated in this study at the Department of Clinical Physiology, Lund University Hospital, Lund, Sweden. The local investigational review board approved the study.
Standardized Physical Activities
We chose a range of physical activities known to cause baseline wander, myoelectric noise, or both. During a single 1-hour study, each subject had ECGs recorded at rest and during performance of the following physical activities:
- holding a 3-kg load with both arms at a 45° angle when supine ("arms up"),
- walking on a treadmill,
- turning slowly from supine to the left side,
- turning slowly from supine to the right side, and
- cycling on an ergometer.
ECG Acquisition
For each subject, 2 simultaneous recordings of the ML 12-lead ECG and the EASI ECG were made for each type of physical activity. The electrode type used was Niko•UNI•4 silver/silver-chloride electrodes (Niko Medical Products, Rødovre, Denmark); a Page Writer XLi (Philips Medical, Oxnard, Calif) acquired the ECG signal for 10 seconds and stored it digitally for later offline processing.
Figure 1A
shows the 4 recording electrode positions of the EASI system. Positions A, E, and I are those of the Frank vectorcardiographic system.10 E is placed on the lower extreme of the body of the sternum. A and I are placed on the left and right midaxillary lines, respectively, in the same transverse plane as E. S is placed on the sternal manubrium. We used the same ground electrode for both the ML and the EASI system in our study.
| 12-Lead ECGs with limb leads are particularly susceptible to muscle movement artifact.
|
The 3 bipolar quasi-orthogonal leads of the EASI system were generated offline by pairwise subtraction of signals recorded at the E, A, S, and I sites. Lead AI views the electrical activity of the heart in a left-to-right direction. Lead AS views the electrical activity of the heart both in left-to-right and caudal-cranial directions but also contains a small anterior-posterior component. Lead ES views the electrical activity of the heart in a caudal-cranial direction and also contains a considerable anterior-posterior component.
Optimal algebraic transfer coefficients have been determined for generating a derived 12-lead ECG from the ECG recorded at the EASI positions.11
| The ML ECG systems use 6 precordial electrodes and 4 torso sites for limb electrodes.
|
Figure 1B
shows the lead placement for the ML modification of the standard 12-lead ECG. The right arm and the left arm electrodes are positioned in the right infraclavicular fossa and left infraclavicular fossa, respectively, medial to the border of the deltoid muscle and 2 cm below the lower border of the clavicle. The left leg electrode is positioned in the anterior axillary line, halfway between the costal margin and the crest of the ilium.
Noise Measurements
Two types of noise measurements were computed for each of the leads in both ECGs (Figure 2). These measures reflect the level of baseline wander and myoelectric noise, respectively. Computation of these measurements requires software detection of the QRS complexes.12
View larger version (17K):
[in this window]
[in a new window]
|
Figure 2 A, Original 10-second epoch in a precordial lead during a change in body position. B, Estimated baseline wander during the epoch. C, Filtered-out myoelectric noise during the epoch. A segment around and including each QRS complex has been blanked.
|
|
Baseline wander is due to variations in electrode impedance caused by perspiration, respiration, and physical activity. Such low-frequency noise was quantified by fitting a regression model to the original ECG signal by using a regularized least-squares criterion.13 The method has time-varying properties that are essential in avoiding distortion at the signal end points; such distortion is otherwise a serious problem when time-invariant, linear filtering is used on short-duration signals. Baseline wander was defined as the peak-to-peak amplitude of the estimated baseline during the 10-second recording (Figure 2B).
Noise due to myoelectric activity primarily contains frequency components greater than 15 Hz. This type of activity was quantified from a signal that had been subjected to high-pass filtering (by using a filter cutoff frequency of 15 Hz) and subsequent blanking of the QRS complexes in an interval of 200 ms centered on the R wave. Finally, the root mean square value was computed in each R-R interval of the processed signal, and the largest value was taken as the final measurement of myoelectric noise (Figure 2C).
The length of the R-R interval for noise measurements is sufficient for all physiological heart rates. The influence of large-amplitude T waves in the measurement interval is avoided by using high-pass filtering.
Data Analysis
For each lead in each subject, we calculated the noise measurements at rest and during each physical activity for both the ML and the EASI-derived 12-lead ECGs. If these measures were within 25 µV for the myoelectric noise or within 300 µV for the baseline wander, results were declared "equal" for the 2 systems. If the difference was greater than these respective thresholds, a "winner" was declared. These 2 threshold values were selected with reference to a signal amplitude at which the noise begins to have a clinical impact on the ECG interpretation. We used the 
2 statistic to compare the proportion of winners in the EASI and ML lead placement groups. In situations in which no winners occurred in 1 of the 2 groups, the Fisher exact test was used. One P value was determined for the summed limb leads and the summed precordial leads for each physical activity separately.
|
Results
|
|---|
Characteristics of the Sample
In 1 of the 20 subjects, 4 of the 6 recordings had such excessive noise that analysis with our method was not meaningful, and therefore that subject was excluded from the study. Table 1 gives demographic descriptions of the remaining 19 study subjects, all of whom were white. The ages ranged from 30 to 64 years (mean, 46 years) and the heights and weights ranged from 1.61 to 1.77 m (mean 1.69 m) and from 54 to 95 kg (mean, 71 kg).
Figure 3 and Table 2 present the outcome of comparisons of baseline wander in the 2 recording systems. We found few differences either at rest or during physical activities required for arms up or cycling. For treadmill exercise and turning supine to left, the EASI-derived 12-lead ECG is superior, attaining a significantly higher number of "wins" in the precordial leads. However, the ML ECG attained a significantly higher number of wins for turning supine to right in both the limb and precordial leads.
View larger version (26K):
[in this window]
[in a new window]
|
Figure 3 Comparison of the outcomes for baseline wander at rest (A) and all 5 physical activities (B-F). On the horizontal axis, the zero indicates equivalent performance of the Mason-Likar (ML) 12-lead and EASI-derived 12-lead electrocardiograms. A longer bar to the right than to the left indicates superiority of the EASI-derived 12-lead electrocardiogram and vice versa. The numbers on the horizontal axis indicate the frequency of these occurrences.
|
|
| The EASI system uses 5 torso electrodes and mathematical transformations to derive a 12-lead ECG.
|
Figure 4 and Table 3 present the outcome of comparisons of myoelectric noise content in the 2 recording systems. We found few differences either at rest or during physical activities required for arms up. For treadmill exercise, cycling, and for turning supine to left, the EASI-derived 12-lead ECG is superior; attaining a significantly higher number of wins in the limb leads. However, the ML ECG attained a significantly higher number of wins for turning supine to right in the precordial leads.
View larger version (25K):
[in this window]
[in a new window]
|
Figure 4 Comparison of the outcomes for myoelectric noise at rest (A) and all 5 physical activities (B-F). On the horizontal axis, the zero indicates equivalent performance of the Mason-Likar (ML) 12-lead and EASI-derived 12-lead electrocardiograms. A longer bar to the right than to the left indicates superiority of the EASI-derived 12-lead electrocardiogram and vice versa. The numbers on the horizontal axis indicate the frequency of these occurrences.
|
|
|
Discussion
|
|---|
Real-time ST monitoring is feasible and accurate14 and can be useful for supporting clinical decisions. However, false alarms are an important issue that must be addressed. Changes in body position may cause ST-T and QRS changes and may trigger false alarms.15–17 The lower the susceptibility of a method to artifacts, the higher is the quality of the signal and the lower the prevalence of false alarms. A previous study18 confirmed that ECG monitors in intensive care units generate a great number of false alarms. As indicated in the literature (and corroborated by vast clinical experience), abrupt changes in the appearance of the ECG due to a change in body position can be erroneously interpreted as an acute ischemic event19 or an episode of arrhythmia. If false alarms occur frequently, staff might disable the alarm, potentially jeopardizing patients’ safety.
The ML system for placement of electrodes allows high-quality recording of 12-lead ECGs during exercise as well as at rest, and the precordial leads of the 12-lead ECGs closely resemble the standard 12-lead ECGs with distal positions for the limb electrodes. However, in many clinical situations, an ECG-recording system that has the same advantages as ML but also has a reduced electrode set would be valuable. Fewer electrodes on more easily located sites would make data acquisition simpler and more rapid for both patients and staff members. The electrode positions of the E, A, S, and I electrodes are easily located, potentially reducing the risk of improperly placed electrodes. With fewer electrodes, the interference with other clinical procedures would be lessened, movement artifacts would be lessened, and patients’ comfort would be increased.
We found no large differences between the ML and EASI systems in susceptibility to baseline wander, most likely because variations in electrode impedance, the primary cause of baseline wander, are due primarily to such factors as perspiration, respiration, and physical activities, which affect the electrodes of the 2 systems equally. The EASI system is, however, less susceptible to myoelectric noise than is the ML system. The most likely reason is the location of the electrodes; the sites of the electrodes of the EASI system are probably not as susceptible to skeletal muscular artifacts as are those of the ML system. The EASI system is much more susceptible to both baseline wander and myoelectric noise than is the ML system when patients turn from supine to the right side, because only the EASI system uses an electrode placed in the right-midaxillary line (Figure 1).
Limitations
One limitation of our study is that some artifact-generating movements such as combing hair, brushing teeth, and shaving, which consequently may cause false alarms, were not included in the protocol. Another limitation is that the subjects were all relatively young (30–64 years old) and healthy, characteristics not generally representative of the population monitored in hospitals. The myoelectric noise and baseline wander might be influenced by age and infirmity.
|
Conclusion
|
|---|
The EASI system is less susceptible to myoelectric noise than is the ML system. However, the systems did not differ significantly with respect to susceptibility to baseline wander. These results, together with those of previous studies, make the EASI system an alternative to the ML system for both in-hospital and ambulatory 12-lead ECG monitoring.
|
ACKNOWLEDGMENT
|
|---|
This study was supported by Philips Medical Systems (Oxnard, Calif).
To purchase reprints, contact The InnoVision Group, 101 Columbia, Aliso Viejo, CA 92656. Phone, (800) 809-2273 or (949) 362-2050 (ext 532); fax, (949) 362-2049; e-mail, reprints{at}aacn.org.
|
REFERENCES
|
|---|
- Mason RE, Likar I. A new system of multiple-lead exercise electrocardiography. Am Heart J. 1966;71:196–205.[Medline]
- Dower GE. EASI 12-Lead Electrocardiography. Roberts, Wash: Totemite Inc; 1996.
- Klein MD, Key-Brothers I, Feldman CL. Can the vectorcardiographically derived EASI ECG be a suitable surrogate for the standard ECG in selected circumstances? In: Computers in Cardiology. Piscataway, NJ: IEEE Computer Society Press; 1997:721–724.
- Drew BJ, Pelter MM, Wung SF, et al. Accuracy of the EASI 12-lead electrocardiogram compared to the standard 12-lead electrocardiogram for diagnosing multiple cardiac abnormalities. J Electrocardiol. 1999; 32(suppl):38–47.
- Dower GE, Yakush A, Nazzal SB, Jutzy RE, Ruiz CE. Deriving the 12-lead electrocardiogram from four (EASI) electrodes. J Electrocardiol. 1988;21(suppl):S182–S187.
- Pahlm O, Pettersson J, Thulin A, Feldman CL, Feild DQ, Wagner GS. Comparison of waveforms in conventional 12-lead ECGs and those derived from EASI® leads in children. J Electrocardiol. 2003;36:25–31.[Medline]
- Drew BJ, Scheinman MM, Evans GT. Comparison of a vectorcardiographically derived 12-lead electrocardiogram with the conventional electrocardiogram during wide complex tachycardia and its potential application for continuous bedside monitoring. Am J Cardiol. 1992; 69:612–618.[Medline]
- Horá
ek BM, Warren JW, Stovicek P, Feldman CL. Diagnostic accuracy of derived versus standard 12-lead electrocardiograms. J Electrocardiol. 2000;33(suppl):155–160.
- Feldman CL, MacCallum G, Hartley LH. Comparison of the standard ECG with the EASI cardiogram for ischemia detection during exercise monitoring. In: Computers in Cardiology 1997. Piscataway, NJ: IEEE Computer Society Press; 1997:343–345.
- Frank E. An accurate, clinically practical system for spatial vectorcardiography. Circulation. 1956;13:737–749.[Medline]
- Feild DQ, Feldman CL, Horác ek BM. Improved EASI coefficients: their derivation, values and performance. J Electrocardiol. 2002;35(suppl):23–33.
- Lagerholm M, Peterson C, Braccini G, Edenbrandt L, Sörnmo L. Clustering ECG complexes using Hermite functions and self-organizing maps. IEEE Trans Biomed Eng. 2000;47:838–848.[Medline]
- Tarvainen MP, Ranta-aho PO, Karjalainen PA. An advanced detrending method with application to HRV analysis. IEEE Trans Biomed Eng. 2002;49:172–175.[Medline]
- Johanson P, Rossberg J, Dellborg M. Continuous ST monitoring: a bedside instrument? A report from the Assessment of the Safety of a New Thrombolytic (ASSENT 2) ST monitoring substudy. Am Heart J. 2001;142: 58–62.[Medline]
- Jensen SM, Haggmark S, Johansson G, Naslund U. Online computerized vectorcardiography: influence of body position, heart rate, radiographic contrast fluid and myocardial ischemia. Cardiology. 1997;88:576–584.[Medline]
- Norgaard BL, Rasmussen BM, Dellborg M, Thygesen K. Temporal and positional variability of the ST segment during continuous vectorcardiography monitoring in healthy subjects. J Electrocardiol. 1999;32:149–158.[Medline]
- Norgaard BL, Rasmussen BM, Dellborg M, Thygesen K. Positional changes of spatial QRS- and ST-segment variables in normal subjects: implications for continuous vectorcardiography monitoring during myocardial ischemia. J Electrocardiol. 2000;33:23–30.[Medline]
- Chambrin MC, Ravaux P, Calvelo-Aros D, Jaborska A, Chopin C, Boniface B. Multicentric study of monitoring alarms in the adult intensive care unit (ICU): a descriptive analysis. Intensive Care Med. 1999;25:1350–1352.[Medline]
- Drew BJ, Adams MG. Clinical consequences of ST-segment changes caused by body position mimicking transient myocardial ischemia: hazards of ST-segment monitoring? J Electrocardiol. 2001;34:261–264.[Medline]
Top
Abstract
Methods
