THE EFFECT OF LOCATION OF LIMB ELECTRODES ON 12-LEAD ECG AND BODY SURFACE POTENTIAL MAPPING

M. Stenroos^{1, 2}, K. Simelius^{1, 3}, I. Tierala², J. Nenonen¹, L. Toivonen², T. Katila¹
¹Laboratory of Biomedical Engineering, P.O. Box 2200, FIN-02015 HUT, Finland
²Helsinki University Central Hospital, Division of Cardiology and Biomag laboratory,
P.O. Box 340, FIN-00029 HUS, Finland
³Nokia Corporation, P.O. Box 88, FIN-33720 Tampere, Finland

Abstract: We studied the effect of limb electrode placement on the 12-lead ECG and body surface potential mapping (BSPM) with a 128-channel BSPM system. Four subjects were measured. A total of 60 electrodes were attached to the arms and the left leg, the rest of the electrodes were attached to the chest and the back. 8000 different Wilson central terminals (WCT) and 400 different bipolar limb lead combinations (I, II, and III respectively) were calculated from the data. The analysis methods used included visual inspection, heart vector frontal angle calculations (the 12-lead ECG), and correlation coefficient calculations between leads measured against different WCTs.

It was found that the location of the limb electrodes has a significant effect on the limb leads of the 12-lead ECG, when the upper limb electrodes are close to the heart. The unipolar chest leads and the most of the BSPM leads were not sensitive to the chosen WCT (R > 0.95). However, the chosen WCT had a remarkable effect on the most distal BSPM leads. When body surface potential maps were inspected as spatial potential distributions, the chosen WCT had only an offset effect on the data.

INTRODUCTION

In electrocardiography (ECG) the potentials are always measured against a reference. The standard 12-lead ECG consists of 3 bipolar and 3 unipolar limb leads, and 6 unipolar chest leads [1]. In the bipolar limb leads the voltage is measured between two limb electrodes:

, (1)

and in the unipolar leads (2) the reference potential is the average of the potentials of the three limb electrodes (Ф_L, Ф_R, and Ф_F). This standard reference is called the Wilson Central Terminal (WCT) (3).

(2)

(3)

Also in body surface potential mapping (BSPM), WCT is usually the reference potential. As all the BSPM leads have the same reference, the differences between instantaneous body surface potential maps measured against different references are of offset nature only. However, as WCT is a function of time, differences in maps may occur when sequential maps are subtracted from each other (time difference maps).

The limb electrodes are normally placed in distal positions, but e.g. in pedal stress measurements and during surgery non-distal locations (shoulders, hips) are used. The effect of the location of the limb electrodes on QRS waveforms has been studied before by Pahlm et al. [2]. However, we measured more electrode configurations and analyzed also P and T waves, and body surface potential maps.

METHODS

We studied the effect of the limb electrode placement on the ECG and BSPM with a 128-channel BSPM system [3]. Four subjects were measured. A total of 60 electrodes were attached to the arms and the left leg; the electrodes on each limb were in 5 equidistally spaced zones. The first zone was around the shoulders and hips, and the last on the wrists and the left ankle. The rest of the electrodes were attached to the chest and the back to provide the V1-V6 chest leads and body surface potential maps. The long-term average of the amplifier common, equalized to patient common by a driven right leg feedback loop, was used as the electrical reference.

8000 different WCTs and 400 different bipolar limb lead combinations (I, II, and III respectively) were calculated from the data. The 12-lead ECGs were both inspected visually, and by calculating heart vector frontal axis orientations. When large differences in lead plots were found, also respective WCT differences were studied; this reflects the difference in ECG chest leads and body surface potential maps. Also correlation coefficients were calculated.

RESULTS

It was found that the location of the limb electrodes has a significant effect on the 12-lead ECG, when the upper limb electrodes are close to the heart. Between more distal positions of L and R, no change was observed in the ECG morphology. Altogether, the largest differences were found when moving the L electrode from anterior to posterior side of the shoulder. Hence the largest changes in the 12-lead ECG occur in leads I, III, and aVL. The QRS complexes of leads I and III (one subject, different electrode combinations) are shown in figure 1. As can be seen from the lead plots, the frontal axis orientation changes significantly. The angles with different electrode locations are listed in table 1.

Figure 1. QRS complexes of leads I and III with different electrode combinations.

TABLE I
QRS frontal axis angles for all the subjects.

	1	2	3	4
Distal	97	90	21	51
Medial	98	89	52	67
Anterior	96	88	40	60
Posterior	71	67	13	19

In the 12-lead ECG chest leads the WCT-induced differences are considerably smaller than in the limb leads. The morphology doesn’t change significantly; only some minor scaling differences occur. The largest absolute differences were found in R and S peaks, and the largest relative differences in T wave peaks. The same applies to most of the BSPM leads. In chest leads close to the heart correlation coefficients were between 0.95 and 0.99. However, in the most distal BSPM electrodes (e.g. on the right side near hips) the measured voltages are no larger than the largest WCT differences (see Fig. 2). In these leads the chosen WCT changes the data significantly (correlations < 0.3 or negative).

However, usually the BSPM data are inspected with aid of 2D or 3D visualizations of potential distributions, and there the different WCTs are noticed as offsets only (correlation > 0.999). Unlike the frontal axis angle in the 12-lead ECG, the direction of the minimum-maximum vector of the potential map is WCT-independent.

Figure 2. WCT-induced QRS differences in unipolar leads (measured against distal WCT).

DISCUSSION

The results show that in the 12-lead ECG the limb leads are sensitive to electrode locations. When proximal electrode locations are used, the electrodes should be placed close to the medial plane, and rather on anterior than posterior side of the shoulders, and as far away from the heart as possible. The chest leads are practically insensitive to the chosen WCT.

Body surface potential maps are not sensitive to the reference potential. However, if single leads are studied in time domain, the morphology of some small-amplitude leads changes significantly. Noticeable is that the minimum-maximum vector is WCT-insensitive, whereas the frontal axis angle in 12-lead ECG is very sensitive to the electrode locations.

REFERENCES

[1] P.W. MacFarlane, “Lead Systems”, in Comprehensive Electrocardiography, pp. 315-352. New York: Pergamon Press, 1989.

[2] O. Pahlm, W.K. Haisty Jr, L. Edenbrandt, et al.. ”Evaluation of changes in standard electrocardiographic QRS waveforms recorded from activity-compatible proximal limb lead positions,” Am J Cardiol, vol. 69, pp. 253-257, 1992.

[3] K. Simelius, I. Tierala, T. Jokiniemi, et al.. ”A body surface potential mapping system in clinical use,” Med & Biol Eng & Comput, vol. 34, pp. 107-108, 1996.