A Proposal for a 9‐Lead Electrocardiogram Recorded via the Wilson's Central Terminal

Abstract

The author argues that the universally employed 12‐lead standard electrocardiogram (ECG) which consists of a conglomerate of 3 bipolar limb leads, 6 quasi‐unipolar precordial leads acquired via the Wilson's central terminal (WCT), and 3 augmented limb leads recorded via the Goldberger's changing assembly of limb connections, different for each of the augmented leads, is scientifically unsuitable and can be replaced by a 9‐lead ECG format comprising the 6 precordial V₁‐V₆ leads and leads VR, VL, and VF, all recorded via the WCT. The reasons and the advantages of such as switch are being discussed.

The standard 12‐lead electrocardiogram (ECG) as used universally in the past 70 years comprises a configuration with 3 different lead acquisition components: (1) the 3 Einthoven bipolar limb leads (I, II, and III),1 measuring potential differences between pairs of limbs, (2) the 6 “unipolar” precordial leads (V₁‐V₆) via the Wilson's central terminal (WCT),2 and (3) the 3 augmented limb leads (aVR, aVL, and aVF) via the Goldberger's terminal (GT),3 which consists of the average of potentials of left arm and left leg for aVR, right arm and left leg for aVL, and left arm and right arm for aVF (Figure 1A). Following the introduction of leads I, II, and III,1 the quest began to record body potentials from the thorax with the negative electrocardiograph pole connected to a remote site of the body (indifferent electrode) (semi‐direct leads)4, 5 which culminated to the WCT consisting of the connection of electrodes from the right arm, left arm and left leg with the interposition of 3 fixed 5000 Ω resistors.2 This meant to produce an electrode with theoretically a zero potential, so that the leads recorded via this system could be considered truly unipolar; however, this was not feasible in a strict sense, and although the WCT potential is close to zero, it is not zero but negligible, and Wilson knew about it.2 Thus there are no truly unipolar leads, and strictly speaking leads V₁‐V₆ are bipolar leads. The WCT, early in its introduction was used for the acquisition of “unipolar” limb leads (VR, VL, and VF) by connecting the WCT to the negative pole of the electrocardiograph, and the positive pole of the electrocardiograph to the right arm, left arm, and left leg, respectively.2 Such leads were recorded by this author, by repeating the ECG of Fig 1A, and after connecting the lead electrodes for V₁, V₂, and V₃ to the right arm, left arm, and left leg, respectively (Fig. 1B). Goldberger felt that there was no reason to use resistors since the resistance of the skin‐electrode interfaces was high, and thus one could get a central terminal (GCT) by connecting with ordinary electrical wires the right arm, left arm and left leg to the negative pole of the electrocardiograph, and use the lead connected to the positive pole of the electrocardiograph to record the 6 precordial leads V₁‐V₆, and the “unipolar” VR, VL, and VF leads,3 as Wilson did via his WCT.2Dissatisfied with the low amplitude of VR, VL, and VF, he produced another configuration (GT) by connecting 2 of the 3 extremities, in 3 different combinations, excluding each time from his “terminal” the extremity on which the “exploring” positive electrode would be attached, connecting the GT assembly to the negative pole of the electrocardiograph and the positive pole of the electrocardiograph to the limb for which the augmented potential was sought.3 This was Goldberger's innovation, and its motivation was to forgo the expense for construction of the WCT, and to produce 3 limb leads with larger amplitudes (aVR, aVL, and aVF) than the ones recorded with the WCT (VR, VL, VF)! Otherwise Goldberger himself felt that the morphology of the ECG leads recorded via the GT was identical to the one recorded via the WCT3 (Fig. 1A and 1B). The concern about the amplitude of VR, VL, and VF is understandable since the ECG trace was at that time thick,2, 3, 4, 5 and one could not easily discern the morphology of the waves and complexes when the inscribed ECGs were low in amplitude. However subsequent technological advancements provide ECG recordings with thin trace, and the reliable automated measurements of amplitudes and intervals which are currently available upon completion of an ECG recording, eliminates any reason for preferring aVR, aVL, and aVF to VR, VL, and VF. Besides there is a mathematical relationship between the corresponding leads recorded via the CT and the WCT showing that throughout their inscription there is a 50% augmentation of the former as compared to the latter.6

Figure 1.

The patient is a 55‐year‐old man with hypertrophic cardiomyopathy and diabetes mellitus referred to Cardiology for dyspnea on exertion. ECG A is a standard 12‐lead ECG. ECG B is a copy of ECG A, which includes only leads V₁‐V₆. ECG C includes only leads VR, VL, VF, recorded by attaching the lead electrodes V₁, V₂, and V₃ to the right arm, left arm, and left leg, respectively, and was recorded 2′ and 56′′ after the ECG A. ECGs B and C combined constitute the proposed 9‐lead ECG.

The current ECG acquisition configuration resembles an edifice, designed and built by three different architects, working in isolation at different times, concentrating on their part, and simply adding it to what was entrusted to them for completion. Besides, the configuration is unscientific, nonsensible, and nonamenable to measurements and quantitative assessment, and here is why: (1) measuring difference in potentials at 2 different points of the body surface (leads I, II, and III) when one could explore absolute (or close to it) potential at a single body surface point makes no sense; (2) employing a changing “terminal” as the GT is (a terminal should be an invariant point of reference) to record aVR, aVL, and aVF, and for the reasons cited above for which Goldberger constructed it (currently meaningless) makes no sense; (3) The artificiality of the 50% augmentation of leads aVR, aVL, and aVF, mixed with the “unipolar” V₁‐V₆ leads (the ones which make most sense), and the bipolar leads I, II, and III are a strange mélange, not amenable to scientific conceptualization, analysis, and presentation.

Taking into consideration the mathematical relationship between aVR, aVL, and aVF and VR, VL, and VF,6 there is nothing that is accomplished by the former set of leads that cannot be achieved by the latter. Also considering the mathematical relationship between the leads I, II, and III and aVR, aVL, and aVF,6 and by extension VR, VL, and VF, there is nothing that is accomplished by the first and second sets of leads that cannot be achieved by the third. Consequently a 9‐lead ECG comprising leads VR, VL, VF, and V₁‐V₆ (Fig.1B) appears to be more scientifically grounded, than the currently used 12 lead ECG (Fig 1A). A 9‐lead ECG configuration can also provide additional 2.5 s worth of data (total 12.5 s), instead of the 10 s worth of data provided with the 12‐lead ECG in the electronically based or hardcopy printed ECGs, due to space freed in both these 2 formats by the elimination of leads I, II, and III (Fig 1B). One of leads VR, V₁ or V₆ could be adopted as the “arrhythmia lead” at the bottom of the ECG display. This author favors lead VR, considering the avalanche of useful applications for lead aVR (formerly considered a “useless lead”) which is recently reported. The proposed 9‐lead ECG will not provide of course any new, or better, or more accurate, information, and it will not be less expensive than the method in current use. Calculation of the frontal plane axis could be somewhat affected by the proposed system by eliminating the “useful” redundancy provided by leads I, II, and III, but as soon as readers are familiarized with the VR, VL, and VF leads, crude estimation (which suffices for practical purposes) of the frontal QRS axis of complexes and waves will be easy; beside contemporary ECG management systems provide both frontal and horizontal P, QRS, and T axes automatically upon acquisition of the ECG, and clinicians rely on such automated measurements. For example the automated P wave, QRS complex, and T wave frontal axes of the ECG of Fig1A are 39⁰, 12⁰, and 173⁰, respectively and manual measurements cannot accomplish such accuracy.

Some counterarguments for the change advocated above should be considered. Identification of right and left axis deviation in the ECG is easier in leads I and II, but this may be as a result of our familiarity with the current ECG lead configuration. Thus, when our frontal axis conceptualization changes to one based on a relationship of the frontal axis to a “Y‐shaped” lead configuration of VR/VL/VF with an angle of 120⁰ between VF and VL, VF and VR, and VR and VL, estimation of the frontal QRS, P, and T axis, could easily be accomplished. Of course abandoning these 2 leads, i.e., I and II, or any lead(s), will not be easy initially, and should not be taken lightly, considering that they are ingrained in our working routines. Intuitively true bipolar ECG leads have some advantage over leads obtained with a WCT reference. Consider for example the case of a single faulty frontal electrode, which could influence all ECG leads recorded via the WCT. At first one hopes that there remain the useful “uncorrupted” set of standard bipolar leads. However this may be an illusion, since a faulty frontal electrode may also “corrupt” the standard bipolar leads. Needless to say that the above are speculations, and what is needed is research into the characterization of the systematic “corruption” imparted on both bipolar leads and the leads recorded via the WCT by one or more faulty frontal electrodes.

Advantages that would be accrued by a switch from the current 12‐lead ECG format to a 9‐lead configuration include: (1) Uniform acquisition of all 9 leads via the WCT; (2) compatibility of ECG format with that used for multilead body surface potential maps via the WCT; (3) compatibility of ECG format with that used for the recording of intracardiac electrograms via WCT during electrophysiologic testing; (4) compatibility of measurements of amplitudes of all ECG waves, complexes, and deviations of segments and intervals, acquired at rest or exercise by a variety of acquisition configurations, since all would be employing the WCT; and (5) generation of diagnostic and prognostic algorithms, which will be physiologically meaningful since their components will have a common reference (the WCT), instead of the ones employing the unscientific mixture of bipolar, “unipolar,” and augmented leads. The physiological meaningfulness of diagnostic and prognostic algorithms is not of course provided only by using a common reference for the ECG leads acquisition, but also by our interpretation of ECG based on pathophysiological and biophysical concepts used for the interpretation. However advantages could be accrued by recording all ECG leads with a common reference. The only impediment for an adoption of a 9‐lead ECG is our reverence for the past and the tradition, our conservatism, the large deposit of literature data using the 12‐lead ECG format, and our reluctance to confront the novel. But is it not a break with the past an inherent prerequisite to the scientific advancement?