Vectorcardiographic diagnostic & prognostic information derived from the 12‐lead electrocardiogram: Historical review and clinical perspective

Vectorcardiographic diagnostic & prognostic information derived from the 12‐lead electrocardiogram: Historical review and clinical perspective

    Vectorcardiographic diagnostic & prognostic information derived from the 12-lead electrocardiogram:historical review and clinical per...

1MB Sizes 0 Downloads 44 Views

    Vectorcardiographic diagnostic & prognostic information derived from the 12-lead electrocardiogram:historical review and clinical perspective Sumche Man MD, Arie C. Maan PhD, Martin J. Schalij MD, Cees A. Swenne PhD PII: DOI: Reference:

S0022-0736(15)00128-4 doi: 10.1016/j.jelectrocard.2015.05.002 YJELC 52053

To appear in:

Journal of Electrocardiology

Please cite this article as: Man Sumche, Maan Arie C., Schalij Martin J., Swenne Cees A., Vectorcardiographic diagnostic & prognostic information derived from the 12-lead electrocardiogram:historical review and clinical perspective, Journal of Electrocardiology (2015), doi: 10.1016/j.jelectrocard.2015.05.002

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

RI P

T

Vectorcardiographic diagnostic & prognostic information derived from the 12‐lead electrocardiogram: historical review and clinical perspective

SC

Sumche Man, MD, Arie C. Maan, PhD, Martin J. Schalij, MD, Cees A. Swenne, PhD

AC

CE

PT

ED

MA NU

Department of Cardiology, Leiden University Medical Center, Leiden, The Netherlands

Correspondence: Cees A. Swenne, PhD Cardiology Department, Leiden University Medical Center PO Box 9600, 2300 RC Leiden, The Netherlands E-mail: [email protected]

ACCEPTED MANUSCRIPT ABSTRACT In the course of time, electrocardiography has assumed several modalities with varying electrode

T

numbers, electrode positions and lead systems. 12-lead electrocardiography and 3-lead

RI P

vectorcardiography have become particularly popular. These modalities developed in parallel through the mid-twentieth century. In the same time interval, the physical concepts underlying electrocardiography

SC

were defined and worked out. In particular, the vector concept (heart vector, lead vector, volume

MA NU

conductor) appeared to be essential to understanding the manifestations of electrical heart activity, both in the 12-lead electrocardiogram (ECG) and in the 3-lead vectorcardiogram (VCG). Not universally appreciated in the clinic, the vectorcardiogram, and with it the vector concept, went out of use. A revival of vectorcardiography started in the 90’s, when VCGs were mathematically synthesized from standard 12lead ECGs. This facilitated combined electrocardiography and vectorcardiography without the need for a

ED

special recording system.

PT

This paper gives an overview of these historical developments, elaborates on the vector concept and

CE

seeks to define where VCG analysis/interpretation can add diagnostic/prognostic value to conventional

AC

12-lead ECG analysis.

ACCEPTED MANUSCRIPT INTRODUCTION Since cardiology emerged from internal medicine as a distinct medical specialty[1], various modern

T

technologies have been prominently present in its diagnostic and interventional/therapeutic procedures.

RI P

One of these was the (surface) electrocardiogram (ECG) that has not subsided since its introduction more than a century ago, and is nowadays present in many forms with a plethora of electrode configurations

SC

and recording durations in settings varying from resting conditions to dynamic situations[2]. When tracing

MA NU

back the development of electrocardiology from its initial primitive form at the end of the nineteenth century (the first human electrocardiogram was published by Waller in 1887[3]) to its current form in the beginning of the twenty-first century, tens of seminal contributions by scientists, engineers and physicians can be mentioned[4,5].

ED

The vectorcardiogram (VCG) is a special form of ECG. In this paper, we discuss the origins and essentials of the VCG and how this emerged in the 1950’s and 60’s partly replacing the standard 12-lead ECG. We also

PT

discuss the temporary decrease in the interest in vectorcardiography in the 1970’s and 80’s, and the revival of vectorcardiography in the 90’s, when it became increasingly popular to mathematically

CE

synthesize a VCG from a standard 12-lead ECG, eliminating the need for dedicated VCG recording

AC

equipment. Finally, we discuss the potential future incorporation of VCG-derived information in routine clinical 12-lead electrocardiography.

ACCEPTED MANUSCRIPT ORIGIN OF THE ELECTROCARDIOGRAM The information content of any particular form of electrocardiography obviously depends on the

T

electrode configuration (number and positions of the electrodes utilized to sample the body surface

RI P

potential distribution in space and time). Current electrocardiography relies in large part on signals derived from 9 electrodes; 3 extremity electrodes at the left arm (LA), right arm (RA) and left foot (LF), as

SC

already used by Einthoven[6,7], and 6 chest electrodes (C1-C6) of which of the positions were standardized in 1938[8-10]. The twelve leads as we are using them today have been introduced by

MA NU

Einthoven (who derived extremity leads I, II, and III from the LA, RA and LF electrodes[6,7]), Goldberger (who derived the augmented extremity leads aVR, aVL, and aVF from the same electrodes[11]), and Wilson et al. (who contributed to the derivation of the precordial leads V1-V6 from chest electrodes C1-C6 by introducing the “central terminal”, i.e., the average of the LA, RA and LF extremity electrode potentials

ED

[12]). To do justice to history, chest electrodes had been used earlier, however in combination with an electrode on the back, by Waller[3] and Wolferth et al [13].

PT

In fact, the term 12-lead ECG is slightly confusing, because nine electrodes can produce only eight

CE

independent leads. As the six extremity leads I, II, III, aVR, aVL and aVF are all derived from three electrodes, only two extremity leads carry independent information, the other four are completely

AC

redundant. Usually, leads I and II are selected to represent the six extremity leads. Hence, extremity leads I and II, and chest leads V1-V6 contain all information obtained with electrodes LA, RA, LF and C1-C6. A variant of the standard 12-lead ECG is recorded with the extremity electrodes placed on the torso just proximal to the right arm, left arm and left leg (Mason-Likar electrode configuration[14]). This variant 12lead ECG, of which the information content is slightly different from the standard 12-lead ECG[15], is typically used in situations where distal extremity electrodes at wrists and ankle would likely cause noise (e.g., in the setting of exercise) or where distal extremity electrodes would be inconvenient (e.g., in the setting of emergency and/or monitoring).

ACCEPTED MANUSCRIPT HEART VECTOR AND LEAD VECTOR Another, parallel line of development in electrocardiography relates to the vector concept. In the first

T

articles concerning the human electrocardiogram,[16,17] Augustus D. Waller pointed out the dipolar

RI P

nature of the cardiac electric generator on the basis of a body surface isopotential map, see Figure 1. Because it is possible to describe the electric generator of the heart reasonably accurately with an

SC

equivalent dipole, it is natural to display it in vector form[18] (in physics, an electrical dipole is denoted as a vector, depicted as an arrow with given orientation and length, representing the direction and strength

MA NU

of the dipole, respectively). Einthoven already displayed the electrical activity of the heart in the form of an arrow (“Pfeil”), on the shaft of which a segment pq was chosen to represent the momentary potential difference E that was projected on the three sides of an equilateral triangle, thus resulting in potential differences e1, e2, and e3 in leads I, II, and III, respectively [6,7], see Figure 2.

ED

An essential step forward was made by Burger[19] and coworkers who laid the theoretical background for vectorcardiography and conceived — based on thorough mathematical and physical knowledge and

PT

ample experimentation with a phantom model — the notions of heart vector, lead vector and image

CE

space[20,21]. In short, the heart vector represents the momentary total cardiac electrical dipole strength and direction. The momentary amplitudes of the voltages in the various electrocardiographic leads are

AC

determined by projecting the heart vector on the lead vectors associated with each electrocardiographic lead, and multiplying these projections by the lead-vector magnitudes (strengths), see Figure 3 for explanation. For a given heart vector H and a given ECG lead with lead vector c , Burgers equation for the voltage in that lead, V, reads

V  c H

(Eq. 1)

Direction and strength of a particular lead vector depend on the distances of the involved electrodes to the heart and the inhomogeneity of the conductive properties in the thorax. These properties represent the direction in which the particular lead “sees” the heart vector, and the sensitivity of this lead, respectively. Thus, lead vectors take the distortions caused by the boundary and internal inhomogeneities

ACCEPTED MANUSCRIPT of the body into account. The lead vector concept is a major improvement from the common oversimplified interpretation in which it is assumed that the spatial direction of the line connecting an electrode pair determines the orientation of the lead. It is essential to realize that a particular ECG lead

RI P

T

“sees” the heart vector from a different direction and with a different sensitivity as what would be expected on the basis of the positions of the electrodes associated with that lead. Burger, realizing that lead vectors mapped the heart vector into a space different from physical space, termed this “image

SC

space” in a seminal 1948 publication[20]. At the same time as the image space was introduced, he

MA NU

described that this would facilitate to inversely reconstruct the heart vector from the ECG.

The difference between Einthoven’s and Burger’s approaches is dramatic, as becomes immediately clear when comparing Einthoven’s equilateral triangle projection (Figure 2) with Burger’s scalene triangle projection with correction for lead strengths (Figure 3). E.g., Einthoven’s lead vector I has a horizontal

ED

direction and its strength equals the lead II and lead III strengths, whereas Burger’s more correct lead

PT

vector I inclines about 17° and its strength is considerably smaller than that of leads II and III[22].

More complex modelling approaches of the electrical source of the ECG (moving instead of fixed dipole,

CE

multiple dipole and multipole[23]) have never gained popularity, and the above described theoretical basis of electrocardiography (equivalent dipole at a fixed location, heart vector, lead vector) is considered

AC

a valid approach today. In clinical practice, these relevant physical considerations are often neglected, however: ECG interpretation is still done in Einthoven’s physical space. E.g., several clinical textbooks postulate that the extremity leads are purely in the frontal plane and the chest leads are purely in the transverse plane; moreover, the angles between the lead vectors in the frontal plane are supposed to be multiples of 30°. Figure 4 shows that these assumptions are far from correct, and the irregular distribution of the directions and strengths of the lead vectors in the 12-lead ECG complicate straightforward ECG interpretation.

ACCEPTED MANUSCRIPT THE VECTORCARDIOGRAM Accepting as model of the electrical activity of the heart a heart vector (the term vector was first used in

T

this respect by Williams in 1914[24]) that represents the instantaneous dipole strength and direction, the

RI P

major goal of electrocardiography would be the dynamic measurement of this heart vector. This would necessitate the recording of a vectorcardiogram (VCG), consisting of three orthonormal leads X, Y and Z,

SC

with lead vectors in the directions of the (orthogonal) main axes of the body and with equal (normalized) lead strengths, thus measuring the dynamic x, y and z components of the heart vector, respectively [22].

MA NU

By combining the xy, xz, yz and xyz amplitudes, two-dimensional or three-dimensional patterns of movement of the heart vector, vector loops[18], can be constructed (see Figure 5) — the name “vectorcardiogram” was chosen by Wilson and Johnston[25] in 1938; earlier, in 1936, Schellong[26] had introduced the word “vectordiagram” while Mann[27], who, in 1920, published the first vector loop,

ED

called it “monocardiogram”. In contrast to scalar representations, these 2D and 3D Lissajous figures give insight into the time relationships between the leads, already mentioned by Williams in 1914[24]. It

PT

becomes immediately apparent from vector loops that the largest amplitudes in the leads are not reached

CE

at the same time. E.g., in Figure 5, this can clearly be seen in the vector loops in the transverse and sagittal planes. In this example, the vector loop in the transverse plane shows that the largest amplitude

AC

in lead Z is attained well before the largest amplitude in lead X, and the vector loop in the sagittal plane shows that the largest amplitude in lead Z is attained well before the largest amplitude in lead Y. Hence, the largest amplitude in the ECG is usually missed, because in general the direction of the heart vector at that very moment is not parallel to one of the lead vectors. The solution in vectorcardiography is to compute, as a fourth scalar lead, the vector magnitude according to the theorem of Pythagoras (rootsum-squared x-y-z amplitudes).

In the next decades, various systems for vectorcardiography were introduced, each with their own electrode configuration. The most known were those proposed by Burger et al.[21], McFee et al.[28], Schmitt et al.[29], and Frank[30]. Of these systems, the Frank system prevailed (see Figure 6). Due to important protagonists of the VCG like Robert P. Grant[31] (1915-1966) and J. Willis Hurst[32] (1920-

ACCEPTED MANUSCRIPT 2011), vectorcardiography was in routine clinical use in various hospitals in the 1960s, but the scalar 12lead ECG, that had half a century of history as a clinical diagnostic tool[33] before the VCG was conceived,

RI P

T

remained the most popular.

DECLINE OF VECTORCARDIOGRAPHY

SC

When both 12-lead electrocardiography and vectorcardiography were in clinical use, several studies appeared that compared ECG and VCG performance. Comparing 12-lead electrocardiography and

MA NU

vectorcardiography (e.g., Frank system) can be done at the level of the information content of the signal, as well as at the level of the diagnostic algorithms[34]. Reasoning in terms of information content, the 12lead ECG is measured by 9 electrodes (right leg reference electrode not included), which means that there are 9−1=8 independent leads. The Frank VCG is measured by 7 electrodes, of which 7−1=6 independent

ED

leads could be constructed. However, conceptually, the electrodes are combined in such a way that 3 leads (the supposedly orthonormal X, Y and Z leads) result. Apart from the question of which of these

PT

electrode configurations comprises the largest information content, it is obvious that the reduction in the

CE

number of leads in the VCG to 3 causes some information loss. It is therefore conceivable that, at the signal level, the 3 VCG leads X, Y and Z contain less information than the 8 ECG leads I, II, V1-V6. The

AC

advantages of the VCG are the orthonormality of the 3 leads and the availability of the phase relationships of these leads, which is a favorable basis for diagnostic algorithms that don’t have access to such data in the 12-lead ECG. In this way, it is understandable that if adequately powerful diagnostic classification procedures are used in extracting diagnostic information[5], the diagnostic information contents (not to be confused with the signal information contents) of the standard 12-lead ECG and the VCG are practically identical[35].

The latter study appeared in 1987, when vectorcardiography had been discontinued at most clinical sites. One of the reasons for this discontinuation may have been the publication by Simonson et al.[36] in 1966. This research was done after several studies (summarized in the same publication[36]) had appeared that claimed superiority of vectorcardiography. Simonson and colleagues provided evidence for the conclusion

ACCEPTED MANUSCRIPT that the diagnostic performance of experienced electrocardiographers was superior for the 12-lead ECG as compared to the VCG. Later, it became evident that automated diagnostic 12-lead ECG algorithms performed at the level of the most accurate cardiologists[37], but for the VCG, statistical programs

RI P

T

performed better than the cardiologists[38]. The diverse conclusions when comparing the diagnostic performance of the ECG to that of the VCG may, hence, well have been caused by the different evaluation methodology (visual inspection by an electrocardiographer versus automated diagnosis by a computer

SC

program). In practice, the standard 12-lead ECG remained the most popular variant and persisted, while

MA NU

the VCG vanished. In the 1967 Recommendations for Standardization of Leads and of Specifications for Instruments in Electrocardiography and Vectorcardiography (see Figure 7), Kossmann and colleagues noted: “That there is redundant electrocardiographic information in the usually recorded 12 leads is generally agreed. However, when the suggestion is made to the clinician that the number of leads he is

ED

using now is excessive, resistance to change is encountered which results probably from the combination of long-ingrained habit and the inadequacy of clinicopathological and other correlations with orthogonal

PT

leads[39]”. This observation was remarkably correct and routine clinical electrocardiography has

CE

continued to exist mainly in the form of the scalar 12-lead ECG.

AC

REVIVAL OF VECTORCARDIOGRAPHY With the decreasing clinical interest for vectorcardiography, attempts were made to mathematically synthesize, by matrix multiplication, a 12-lead ECG from a VCG, thus facilitating those who possessed VCG equipment to perform 12-lead ECG diagnosis, which has led to the development of the “Dower matrix” [40]. The interest in the diagnostic and prognostic value of the VCG has never completely subsided, however, and the past decades have shown a revival of the VCG. In retrospect, the turning point could be situated around 1987-1990 when there were three important publications about matrices for the mathematical synthesis of a Frank VCG from a 12-lead ECG recording — first by Levkov in 1987[41], then by Edenbrandt and Pahlm in 1988 (the “inverse Dower matrix”)[42], and by Kors et al. in 1990[43] — thus facilitating those who possessed standard 12-lead ECG equipment to perform VCG diagnosis and research. This has led to what has been called “12-lead vectorcardiography” [44,45]. This made it possible to choose

ACCEPTED MANUSCRIPT between or even combine 12-lead electrocardiographic and 3-lead vectorcardiographic computer analysis [46].

T

The matrix multiplication that synthesizes a VCG from a 12-lead ECG is algebraically noted as:

RI P

VCG  M  ECG

(Eq. 2)

SC

where VCG is a matrix representation of the computed VCG samples, ECG a matrix representation of the originally recorded 12-lead ECG samples and M a matrix consisting of multiplication factors. This

MA NU

mathematical operation is illustrated in Figure 8; note that only 8 of the 12 ECG leads (I, II, V1-V6) are used because the remaining 4 extremity leads are fully redundant (can straightforwardly be computed from leads I and II and thus contain no independent information). Currently, the matrix by Kors and colleagues [43] is generally accepted as the best method to synthesize a VCG from a 12-lead ECG: “Kors is

ED

still most Frank”[47]. Figure 9 gives an example ECG recording and its synthesized VCG. Note how the Kors matrix coefficients as shown in Figure 8 determine the relative contribution of each original 8 ECG leads to

PT

each of the 3 synthesized VCG leads.

CE

UNIQUE VECTORCARDIOGRAPHY-BOUND INFORMATION

AC

Because the information in the 12-lead ECG is fairly redundant [39-42], reduction of the number of leads from eight (standard 12-lead ECG) to three (VCG) can be achieved with limited information loss [43]. Despite the fact that the VCG actually contains slightly less information than the ECG, VCG analysis/interpretation has additional value compared to ECG analysis/interpretation because the VCG gives access to information that remains unexplored in the standard 12-lead ECG. Examples of such unexplored information are: 

maximal amplitudes of the QRS complex and the T wave (as discussed above, and illustrated in Figure 5, in scalar electrocardiography every lead has its own sensitivity and its own timing of the maximum, while in spatial vectorcardiography there is only one calibrated maximal QRS- or T vector with a specific timing);

ACCEPTED MANUSCRIPT 

QRS- and T-wave axes in three dimensions (instead of an approximated projection in the presumed frontal plane);



spatial QRS- and T-wave integrals which are indexes for dispersion of depolarization and

RI P



T

repolarization, respectively [44];

spatial angle between the QRS- and T-wave axes which is a measure of concordance/discordance of the ECG [45] – see Figure 10;

ventricular gradient (spatial QRS-T integral) which reflects the action potential morphology

SC





MA NU

distribution in the heart[45] – the computation of the ventricular gradient is illustrated in Figure 11; ST vector (oftentimes measured at the J point or 40, 60 or 80 ms thereafter) which assesses ischemia, and is hence also called injury or ischemia vector; 

QRS- and T-loop complexity (healthy hearts have smooth planar loops; compromised hearts have

ED

irregular and distorted loops [46-49]).

Although they are not new, these vectorcardiographic variables have never played a significant role in

PT

clinical vectorcardiography as it was performed several decades ago, although their potential has been

CE

recognized [50]. E.g., Barker, in his book The Unipolar Electrocardiogram (1952)[51], wrote: “At the present time, the determination of the ventricular gradient is not a practical procedure for general use in

AC

electrocardiography. The theoretical foundation upon which it rests, however, are of utmost importance.” All above mentioned variables are rather global VCG characteristics and have been explored to establish normal values, for diagnostic purposes and for risk stratification (see the Appendix for an overview). These new descriptors of abnormality deserve to be further explored and exploited in future electrocardiography, not to replace the standard 12-lead electrocardiogram, but rather to enhance it.

INTEGRATION OF 12-LEAD ELECTROCARDIOGRAPHY AND VECTORCARDIOGRAPHY With mathematically synthesized VCGs, it is feasible to add VCG analysis and interpretation (“12-lead vectorcardiography”[52,53]) to standard 12-lead ECG analysis and interpretation. This leaves the signal acquisition (electrodes, electronics) unaffected, and only requires appropriate additional software to process the ECG signal. Classical VCG diagnostic algorithms can be implemented and executed in addition

ACCEPTED MANUSCRIPT to the current ECG diagnostic algorithms. This will likely increase diagnostic performance, as suggested by Macfarlane and Edenbrandt in 1992[54]. E.g., Kors and colleagues reported that the MEANS ECG/VCG interpretation of the CSE database ECGs yielded a diagnostic accuracy of 69.8% for the ECG alone, a

RI P

T

diagnostic accuracy of 70.5% for the synthesized VCG, and a diagnostic accuracy of 73.6% for the ECG and synthesized VCG in combination [55]. These figures, although impressive, still lag the performance of an expert panel of cardiologists by about 10%. Hence, there is room for further improvement of diagnostic

SC

accuracy. Such improvement could be achieved by utilizing the information in the synthesized VCG that

MA NU

remains unexplored in standard 12-lead ECG analysis. Specifically, so far incompletely explored information in the VCG (QRS integral, T-wave integral, QRS-T integral or ventricular gradient, and QRS-T angle) can help to generate new expertise regarding diagnostics and stratification (see Appendix for examples). Together, this would help to bring clinical electrocardiography to a higher level, and such

AC

CE

PT

ED

expectations are the major driving force for current VCG-oriented research.

ACCEPTED MANUSCRIPT APPENDIX: EXAMPLES OF RESEARCH PROBING VCG-DERIVED INFORMATION To illustrate the concepts discussed in the current paper we have listed in this Appendix briefly overviews

T

of research relating to vectorcardiographic variables as they are enumerated in the section, ”Unique

RI P

vectorcardiography-bound information”. Studies are listed here regard normal values and the potential diagnostic/prognostic relevance of these variables. Investigations have been done using our custom-made

SC

programs for the analysis of the vectorcardiograms (VCGs) synthesized from standard 10-second 12-lead resting electrocardiograms (ECGs) and from longer dynamic 12-lead ECG recordings, e.g., made during

MA NU

exercise tests. These programs are named LEADS [56] and BEATS [57], respectively. The essential difference is that LEADS computes VCG features in one averaged beat, while BEATS computes VCG features in each beat separately, thus allowing for dynamics.

Methodology

ED

Normal limits for the spatial angle between QRS- and T axes (SA) and the ventricular gradient (VG) were

PT

only known for male patients admitted to a hospital but without cardiovascular disease. We determined, in synthesized VCGs, the normal limits (5% and 95%) of SA and VG in a population of 660 male and female

CE

Leiden medicine students aged 18 to 29 years. We observed that SA and VG depend strongly on sex. In female subjects, the SA was more acute than in males (females mean ± SD 66° ± 23°, normal limits 20° and

AC

116°, males mean ± SD 80° ± 24°, normal limits 30° and 130°, P < .001), and the VG magnitude was smaller than in males (females, mean ± SD 81 ± 23 mV∙ms, normal limits 39 and 143 mV∙ms, males mean ± SD 110 ± 29 mV∙ms, normal limits 59 and 187 mV∙ms, P < .001). Note that only angles above the upper normal limit can be called abnormal; however VCGs with angles below the upper normal limit are not by definition normal, e.g., several patients who had SA values below the upper normal limit developed ventricular arrhythmias [58].

Nowadays the Kors ECG-to-VCG transformation matrix is generally preferred above the inverse Dower matrix. Neither spatial angles computed by using the inverse Dower matrix (SA-D) nor by using the Kors matrix (SA-K) had been validated with regard to the spatial angles as directly measured in the Frank VCG

ACCEPTED MANUSCRIPT (SA-F). We computed SAs in 1220 simultaneously recorded 12-lead ECGs and VCGs, in all data, in SA-Fbased tertiles, and after stratification according to pathology or sex. Linear regression of SA-K and SA-D on SA-F yielded offsets of 0.01° and 20.3° and slopes of 0.96 and 0.86, respectively. The bias of SA-K with

RI P

T

respect to SA-F was significantly (P < .001) smaller than the bias of SA-D with respect to SA-F (mean ± SD −3.2° ± 13.9° and 8.0° ± 18.6°, respectively); tertile analysis showed a much more homogeneous behavior of the bias in SA-K than of the bias in SA-D. In pathologic ECGs, there was no significant bias in SA-K; bias

SC

in men and women did not differ. SA-Kors resembled SA-Frank best. Our results show that it is prudent to

MA NU

use the Kors matrix for VCG synthesis. Normal values of the SA and VG computed in Kors-synthesized VCGs are still lacking [59].

Detection of acute ischemia

ED

The ECG is important in the diagnosis and triage of the acute coronary syndrome (ACS), especially in the hyperacute phase, the “golden hours,” during which myocardial salvage possibilities are largest. An

PT

important triaging decision to be taken is whether or not a patient requires primary percutaneous

CE

coronary intervention (PCI), for which, as mentioned in the guidelines, the presence of an ST elevation

AC

(STE) pattern in the ECG is a major criterion.

However, several studies showed that in ACS a completely occluded culprit artery can also occur with a non-ST-elevation (NSTE) ECG. ECGs of 300 ACS patients were vectorcardiographically analyzed. All patients were scheduled for primary PCI and had an angiographically demonstrated completely occluded culprit artery. ECGs were divided in STE and NSTE and it was investigated how this would relate to and be explained by the vectorcardiographically computed ST injury vector. It appeared that 214/300 (71%) of the patients had an STE ECG and 86/300 (29%) had a NSTE ECG (228/72 had single/multivessel disease) prior to catheterization. Injury vector (vectorcardiographically determined magnitude and spatial orientation of the ST amplitude, here computed at the J + 60 ms point) elevation and magnitude were larger in STE than in NSTE patients (32° ± 37° vs. 6° ± 39°, and 304 ± 145 μV vs. 134 ± 72 μV, respectively; P < 0.0001). We conclude that the STE criteria select certain injury vector directions and larger injury vector

ACCEPTED MANUSCRIPT magnitudes. Obviously, with the current ECG criteria, several ACS patients with complete culprit artery occlusions requiring primary PCI do not fulfill these ECG criteria. STENSTE-based ACS stratification needs

RI P

T

therefore further enhancement [60].

Pre-existing non-zero ST amplitudes (diagnostic as well as non-diagnostic) can obscure or even preclude

SC

STE diagnosis. Hence, the potential diagnostic possibilities of ischemia detection by means of changes in

MA NU

the ST vector (ΔS) and changes in the VG (QRST integral) vector (ΔVG) were studied in synthesized VCGs of 84 patients who underwent elective percutaneous transluminal coronary angioplasty. ECG ischemia diagnosis (STE or NSTE), magnitudes and orientations of the ST and VG vectors, and the differences ΔST and ΔVG with the baseline ECG were measured after 3 min of balloon occlusion. The classical criteria

ED

characterized the ECGs of 46/84 (55%) patients after 3 min of occlusion as STE ECGs. A combination of 0.05 mV and 16.2 mV·ms as ΔST and the corresponding ΔVG ischemia thresholds identified 73/84 (87%) of

PT

the patients as ischemic. It was concluded that differential diagnosis by ΔST and ΔVG could dramatically improve ECG-guided detection of patients who urgently require catheter intervention. Development of an

CE

improved algorithm (based on serial analysis of the ST and the ventricular gradient vectors) is needed to

AC

diagnose acute cardiac ischemia [61].

Left ventricular hypertrophy The current electrocardiographic diagnosis of left ventricular hypertrophy (LVH) criteria has low diagnostic accuracy, and in the guidelines, the addition of demographic or anthropomorphic data is suggested as a potential improvement. As hypertrophy affects action potential morphology and intraventricular conduction, QRS prolongation and T-wave morphology may occur and become manifest in the vectorcardiographic variables SA and VG. In this study, we attempted to improve the diagnostic accuracy for LVH by using a combination of demographic, anthropomorphic, ECG, and vectorcardiographic variables. We studied 196 patients who had, on one hand, echocardiographically diagnosed LVH or a normal echocardiogram and, on the other hand, any of the conventional ECG signs for LVH, or normal

ACCEPTED MANUSCRIPT ECGs. We found a discriminant model for LVH diagnosis D = 5.130 × BSA − 0.014 × SA − 8.74 (BSA = body surface area), wherein D greater than or equal to 0 predicts a normal echocardiogram and D less than 0 predicts LVH. The diagnostic accuracy (79%) was better than the diagnostic accuracy of conventional ECG

RI P

T

criteria for LVH (57%). A potential clinical application of this finding is that all ECGs who get either the conventional ECG interpretation “normal” of “LVH” are subjected to the presented algorithm to detect

MA NU

SC

LVH on the basis of SA and BSA[62].

Ventricular arrhythmia Heart failure patients

A low left ventricular ejection fraction is a major criterion for cardioverter-defibrillator implantation for

ED

the primary prevention of serious ventricular arrhythmias in heart failure patients. Because this criterion lacks specificity, other risk indicators for electrical instability were investigated.

PT

T-wave alternans (TWA) is an electrophysiological phenomenon that is associated with serious ventricular

CE

arrhythmias. We compared the performance of TWA amplitude (TWAA) during a complete exercise test with another parameter of electrical instability, the exercise-recovery hysteresis in the ventricular

AC

gradient (VG), in a small group of 34 heart failure patients with implanted cardioverter-defibrillators (ICDs) for primary prevention, half of whom (cases) had and half of whom (controls) had no ventricular arrhythmia during follow-up. TWAA (maximum over the complete test) and the exercise-recovery hysteresis in the VG were computed in ECGs recorded during exercise tests. Receiver operating characteristics (ROC) analysis showed that VG hysteresis discriminated cases and controls with 94.1% sensitivity and 41.2% specificity; hazard ratio was 3.34 (1.17-9.55). ROC analysis of TWAA discriminated cases and controls with 93.8% sensitivity and 23.5% specificity; hazard ratio was 2.07 (0.54-7.91). Hence, VG hysteresis bears predictive potential for arrhythmias in heart failure patients with an ICD for primary prevention, whereas TWA analysis seems to have lesser predictive value in our pilot group. The VG hysteresis is relatively robust for noise, and, as it rests on different electrophysiologic properties than

ACCEPTED MANUSCRIPT TWA, it may convey additional information. Hence, joint analysis of TWA and VG hysteresis may, possibly,

RI P

T

improve the noninvasive identification of high-risk patients [63].

TWA is usually detected in selected ECG leads. TWA amplitude measured in the 12-standard and the 3orthogonal (vectorcardiographic) leads were compared to identify which lead system yields a more

SC

adequate detection of TWA as a noninvasive marker for cardiac vulnerability to ventricular arrhythmias.

MA NU

An adaptive match filter (AMF) technique was applied to the exercise ECG tracings from 58 primary prevention ICD patients, 29 of which had ventricular tachycardia or fibrillation during follow-up (cases), while the remaining 29 were used as controls. Two kinds of TWA indexes were considered: the single-lead indexes, defined as the mean TWA amplitude over each lead (MTWAA), and the lead-system indexes,

ED

defined as the mean and the maximum MTWAA values over the standard leads and over the orthogonal leads. Significantly (P < 0.05) higher TWA in the cases versus controls was identified only occasionally by

PT

the single-lead indexes (odds ratio: 1.0–9.9, sensitivity: 24–76%, specificity: 76–86%), and consistently by the lead-system indexes (odds ratio: 4.5–8.3, sensitivity: 57–72%, specificity: 76%). The latter indexes also

CE

showed a significant correlation (0.65–0.83) between standard and orthogonal leads. Hence, when using the AMF, TWA should be detected in all leads of a system to compute the lead-system indexes, which

AC

provide a more reliable TWA identification than single-lead indexes, and a better discrimination of patients at increased risk of cardiac instability. The standard and the orthogonal leads can be considered equivalent for TWA identification, hence TWA analysis can be limited to one lead system [64].

The value of the SA for prediction of ICD therapy in primary prevention patients with ischemic heart disease was investigated. ICD patients (n=412, 361 men; age, 63±11 years) with ischemic heart disease and a left ventricular ejection fraction <40% were included. After device implantation, the occurrence of appropriate ICD therapy was noted. Survival analysis was performed comparing patients with a SA ≤100° (n=56, 14%) and patients with a SA >100° before implantation. Patients with a SA >100°, the adjusted

ACCEPTED MANUSCRIPT hazard ratio for the occurrence of appropriate device therapy was 7.3 (95% CI, 1.0 to 53.8). Furthermore, a SA ≤100° exhibited a positive predictive value of 98% (95% CI, 95 to 100) for the prediction of an appropriate therapy-free follow-up. A wide SA is a strong predictor of appropriate device therapy in

RI P

T

primary prevention ICD recipients with ischemic heart disease. ICD treatment, although currently

SC

indicated, should be reconsidered in patients who have a SA ≤100° [65].

Dialysis patients

MA NU

Dialysis patients have poor prognosis (annual mortality of about 20%). The spatial QRS-T angle might be useful to identify a subgroup with a high mortality risk. All patients who initiated dialysis therapy between 2002 and 2009 in the academic hospitals of Leiden (LUMC) and Amsterdam (AMC) who were on dialysis for at least 3 months were included. SA was calculated and its relationship with mortality was assessed.

ED

An abnormal spatial QRS-T angle was defined as ≥130° in men and ≥116° in women. In total, 277 consecutive patients (172 male, mean age 56.3+17.0) were included. An abnormal spatial QRS-T angle

PT

was associated with a higher risk of sudden cardiac death (HR 2.99; 95% CI 1.04–8.60). Furthermore, an

CE

abnormal spatial QRS-T angle was of incremental prognostic value, when added to a risk model consisting of known risk factors. Therefore, in chronic dialysis patients the spatial QRS-T angle is a significant and

AC

independent predictor of sudden cardiac death [66].

Pulmonary arterial hypertension ECG interpretation in the setting of pulmonary arterial hypertension (PAH) mainly rests on the VG. The VG expresses asymmetry in the distribution of the action potential morphology in the heart. Because the heart is asymmetrical and myocardial action potentials are not the same everywhere in the heart, the VG is, by definition, different from zero. Processes that locally affect action potentials change the amount of asymmetry and with it the VG. Pulmonary arterial hypertension, no matter whether it is long standing or de novo, changes the VG. The associated phenomenon of right ventricular hypertrophy (RVH) takes time to develop and does not arise in all PAH patients. The counterpart is the fact that there are also patients

ACCEPTED MANUSCRIPT with RVH who do not have PAH. PAH and RVH have different expressions in the ECG [67]. Because of the limited association between hypertension and hypertrophy, it is important to know these typical expressions, and to focus on signs of PAH instead of on signs of RVH when attempts are made to diagnose

RI P

T

PAH in the ECG.

It was demonstrated, in a rat model of PAH, that changes in the VG immediately become apparent when

SC

pulmonary pressure increases [68]. In a population of patients with and without PAH in whom pulmonary arterial pressure (PAP) was measured by right heart catheterization, it was shown that the VG projection

MA NU

in a direction optimized for right-ventricular pressure overload correlated better with PAP than QRS measures for RVH correlate with PAP. On the other hand, in the same population, QRS measures for RVH correlated better with right-ventricular wall thickness than VG correlated with right-ventricular wall

ACKNOWLEDGEMENT

ED

thickness [67].

AC

CE

PT

We thank Dr. Nicolas Davidenko, University of California Santa Cruz, for critically reading our paper.

ACCEPTED MANUSCRIPT REFERENCES [1] Braunwald E. The rise of cardiovascular medicine. Eur Heart J 2012;33:838-45a.

RI P

T

[2] Macfarlane,PW, Van Oosterom A, Pahlm O, Kligfield P, Janse M, and Camm J. Comprehensive Electrocardiology. 2th ed. London: Springer-Verlag; 2011 [3] Waller AD. A Demonstration on Man of Electromotive Changes accompanying the Heart's Beat. J Physiol 1887;8:229-234.

SC

[4] Rautaharju PM. A hundred years of progress in electrocardiography. 1: Early contributions from Waller to Wilson. Can J Cardiol 1987;3:362-374.

MA NU

[5] Rautaharju PM. A hundred years of progress in electrocardiography. 2: The rise and decline of vectorcardiography. Can J Cardiol 1988;4:60-71. [6] EINTHOVEN W, FAHR G, and de Waart A. Über die Richtung und die manifeste Grösse der Potentialschwankungen im menschlichen Herzen und über den Einfluss der Herzlage auf die Form des Elektrokardiogramms. Pflüger Arch ges Physiol 1913;150:275-315.

ED

[7] EINTHOVEN W, FAHR G, and de Waart A. On the direction and manifest size of the variations of potential in the human heart and on the influence of the position of the heart on the form of the electrocardiogram. Am Heart J 1950;40:163-211.

PT

[8] Committee of the American Heart Association for the Standardization of precordial leads. Supplementary report. Am Heart J 1938;15:235-9.

CE

[9] Standardization of precordial leads: Joint recommendations of the American Heart Association and the Cardiac Society of Great Britain and Ireland. Am Heart J 1938;15:107-108.

AC

[10] Standardisation of precordial leads: Joint recommendations of the Cardiac Society of Great Britain and Ireland and the American Heart Association. The Lancet 1938;231:221-222. [11] Goldberger E. A simple, indifferent, electrocardiographic electrode of zero potential and a technique of obtaining augmented, unipolar, extremity leads. Am Heart J 1942;23:483-492. [12] Wilson FN, Johnston FD, Macleod AG, and Barker PS. Electrocardiograms that represent the potential variantions of a single electrode. Am Heart J 1934;9:447-458. [13] Wolferth CC and Wood CW. The Electrocardiographic Diagnosis of Coronary Occulusion By the Use of Chest Leads. Am J Med Sci 1932;183:30-34. [14] Mason RE and Likar I. A new system of multiple-lead exercise electrocardiography. Am Heart J 1966;71:196-205. [15] Man SC, Maan AC, Kim E, Draisma HH, Schalij MJ, van der Wall EE, et al. Reconstruction of standard 12-lead electrocardiograms from 12-lead electrocardiograms recorded with the Mason-Likar electrode configuration. J Electrocardiol 2008;41:211-219. [16] Waller AD. Introductory Address on the Electromotive Properties of the Human Heart. Br Med J 1888;2:751-754.

ACCEPTED MANUSCRIPT [17] Waller AD. On the electromotive changes connected with the beat of the mammalian heart, and of the human heart in particular. Phil Trans R Soc Lond B 1889;180:169-194.

T

[18] Malmivuo J and Plonsey R. Vectorcardiographic Lead Systems. In: Malmivuo J and Plonsey R, editors. Bioelectromagnetism. Principles and Applications of Bioelectric and Biomagnetic Fields, New York, Oxford: Oxford University Press; 1995, p. 290-306

RI P

[19] Van Herpen G. Professor Herman Burger (1893-1965), eminent teacher and scientist, who laid the theoretical foundations of vectorcardiography - and electrocardiography. J Electrocardiol 2013168-174.

SC

[20] Burger HC and VAN MILAAN JB. Heart-vector and leads; geometrical representation. Br Heart J 1948;10:229-233.

MA NU

[21] Burger HC, van BA, and van HG. Heartvector and leads. Am Heart J 1961;61:317-323. [22] Malmivuo J and Plonsey R. Theoretical Methods for Analyzing Volume Sources and Volume Conductors. In: Malmivuo J and Plonsey R, editors. Bioelectromagnetism. Principles and Applications of Bioelectric and Biomagnetic Fields, New York, Oxford: Oxford University Press; 1995, p. 185-226

ED

[23] Malmivuo J and Plonsey R. Summary of the ECG Lead Systems. In: Malmivuo J and Plonsey R, editors. Bioelectromagnetism. Principles and Applications of Bioelectric and Biomagnetic Fields, New York, Oxford: Oxford University Press; 1995, p. 309-312

PT

[24] Williams HB. On the cause of the phase difference frequently observed between homnymous peaks of the electrocardiogram. Am J Physiol 1914;35:292-300. [25] Wilson FN and Johnston FD. The vectorcardiogram. Am Heart J 1938;16:14-28.

CE

[26] Schellong F. Elektrokardiographische Diagnostik der Herzmuskelerkrankungen. Verhandl d deutch Gesellsch f inn Med 1936;48:288-

AC

[27] Mann H. A method of analyzing the electrocardiogram. Arch Intern Med 1920;25:83-94. [28] McFee R and Parungao A. An orthogonal lead system for clinical electrocardiography. Am Heart J 1961;62:93-100. [29] Schmitt OH and Simonson E. Symposium on Electrocardiography and Vectorcardiography: The present status of vectorcardiography. AMA Arch Intern Med 1955;96:574-590. [30] FRANK E. An accurate, clinically practical system for spatial vectorcardiography. Circulation 1956;13:737-749. [31] Frohlich ED. Robert P. Grant, MD. Am J Cardiol 2003;91:646-648. [32] Silverman ME. J. Willis Hurst--a man of achievement. Clin Cardiol 1997;20:584-586. [33] Burch,GE and DePasquale NP. A history of electrocardiography. 1st ed. Chicago: Year Book Medical Publishers; 1964 [34] von der Groeben J. Decision rules in electrocardiography and vectorcardiography. Circulation 1967;36:136-147.

ACCEPTED MANUSCRIPT [35] Willems JL, Lesaffre E, and Pardaens J. Comparison of the classification ability of the electrocardiogram and vectorcardiogram. Am J Cardiol 1987;59:119-124. [36] Simonson E, Tuna N, Okamoto N, and Toshima H. Diagnostic Accuracy of Vectorcardiogram and Electrocardiogram - A Cooperative Study. American Journal of Cardiology 1966;17:829-878.

RI P

T

[37] Willems JL, Abreu-Lima C, Arnaud P, van Bemmel JH, Brohet C, Degani R, et al. The diagnostic performance of computer programs for the interpretation of electrocardiograms. N Engl J Med 1991;325:1767-1773.

SC

[38] Corabian P. Accuracy and reliability of using computerized interpretation of electrocardiograms for routine examinations. 2002;HTA25:14-

MA NU

[39] Kossman CE, Brody DA, Burch GE, Hecht HH, Johnston FD, Kay C, et al. Report of committee on electrocardiography, American Heart Association. Recommendations for standardization of leads and of specifications for instruments in electrocardiography and vectorcardiography. Circulation 1967;35:583-602. [40] Dower GE, Yakush A, Nazzal SB, Jutzy RV, and Ruiz CE. Deriving the 12-lead electrocardiogram from four (EASI) electrodes. J Electrocardiol 1988;21(Suppl):182-187.

ED

[41] Kors JA and van HG. How many electrodes and where? A "poldermodel" for electrocardiography. J Electrocardiol 2002;35 Suppl:7-12.

PT

[42] Nelwan SP, Kors JA, and Meij SH. Minimal lead sets for reconstruction of 12-lead electrocardiograms. J Electrocardiol 2000;33(Suppl):163-166.

CE

[43] Augustyniak P. On the equivalence of the 12-Lead ECG and the VCG representations of the cardiac electrical activity. 10th International Conference on System-Modelling-Control Zakopane 2001:51-56. [44] Van Oosterom A. Reflections on the T waves. Advances in Electrocardiology 2005:807-815.

AC

[45] Draisma HH, Schalij MJ, van der Wall EE, and Swenne CA. Elucidation of the spatial ventricular gradient and its link with dispersion of repolarization. Heart Rhythm 2006;3:1092-99. [46] De Ambroggi L, Aime E, Ceriotti C, Rovida M, and Negroni S. Mapping of ventricular repolarization potentials in patients with arrhythmogenic right ventricular dysplasia: principal component analysis of the ST-T waves. Circulation 1997;96:4314-4318. [47] Perkiomaki JS, Hyytinen-Oinas M, Karsikas M, Seppanen T, Hnatkova K, Malik M, et al. Usefulness of T-wave loop and QRS complex loop to predict mortality after acute myocardial infarction. Am J Cardiol 2006;97:353-360. [48] Priori SG, Mortara DW, Napolitano C, Diehl L, Paganini V, Cantu F, et al. Evaluation of the spatial aspects of T-wave complexity in the long-QT syndrome. Circulation 1997;96:3006-3012. [49] Zabel M, Malik M, Hnatkova K, Papademetriou V, Pittaras A, Fletcher RD, et al. Analysis of Twave morphology from the 12-lead electrocardiogram for prediction of long-term prognosis in male US veterans. Circulation 2002;105:1066-1070. [50] Hurst JW. Thoughts about the ventricular gradient and its current clinical use (Part I of II). Clin Cardiol 2005;28:175-180.

ACCEPTED MANUSCRIPT [51] Barker,JM. The Unipolar Electrocardiogram: A Clinical Interpretation. 1st ed. New York: Appleton-Century-Crofts; 1952

T

[52] Macfarlane PW. Derived Vectorcardiogram - 12 Lead Vectorcardiography. In: Macfarlane PW, Van Oosterom A, Pahlm O, Kligfield P, Janse M, and Camm J, editors. Comprehensive Electrocardiology, London: Springer-Verlag; 2011, p. 404-405

RI P

[53] Macfarlane PW and Pahlm O. 12 Lead Vectorcardiography. In: Macfarlane PW, Van Oosterom A, Pahlm O, Kligfield P, Janse M, and Camm J, editors. Comprehensive Electrocardiology, London: Springer-Verlag; 2011, p. 1951-2006

SC

[54] Macfarlane PW and Edenbrandt L. 12-lead vectorcardiography in ischemic heart disease. J Electrocardiol 1992;24 Suppl:188-193.

MA NU

[55] Kors JA, van HG, Willems JL, and van Bemmel JH. Improvement of automated electrocardiographic diagnosis by combination of computer interpretations of the electrocardiogram and vectorcardiogram. Am J Cardiol 1992;70:96-99. [56] Draisma HH, Swenne CA, van de Vooren H, Maan AC, Hooft van Huysduynen B, van der Wall EE, et al. LEADS: An interactive research oriented ECG/VCG analysis system. Computing in Cardiology 2005;32:515-518. Online available at http://cinc.org/archives/2005/pdf/0515.pdf.

ED

[57] Man S, Maan AC, van der Wall EE, Schalij MJ, and Swenne CA. BEATS: An Interactive Research Oriented Extended ECG Analysis System. Computing in Cardiology 2010;37:1007. Online available at http://cinc.org/archives/2010/pdf/1007.pdf-

PT

[58] Scherptong RW, Henkens IR, Man SC, Le CS, Vliegen HW, Draisma HH, et al. Normal limits of the spatial QRS-T angle and ventricular gradient in 12-lead electrocardiograms of young adults: dependence on sex and heart rate. J Electrocardiol 2008;41:648-655.

AC

CE

[59] Schreurs CA, Algra AM, Man SC, Cannegieter SC, van der Wall EE, Schalij MJ, et al. The spatial QRS-T angle in the Frank vectorcardiogram: accuracy of estimates derived from the 12-lead electrocardiogram. J Electrocardiol 2010;43:294-301. [60] Man S, Rahmattulla C, Maan AC, van der Putten NH, Dijk WA, van Zwet EW, et al. Acute coronary syndrome with a totally occluded culprit artery: relation of the ST injury vector with ST-elevation and non-ST elevation ECGs. J Electrocardiol 2014;47:183-190. [61] Ter Haar CC, Maan AC, Warren SG, Ringborn M, Horacek BM, Schalij MJ, et al. Difference vectors to describe dynamics of the ST segment and the ventricular gradient in acute ischemia. J Electrocardiol 2013;46:302-311. [62] Man S, Rahmattulla C, Maan AC, Holman E, Bax JJ, van der Wall EE, et al. Role of the vectorcardiogram-derived spatial QRS-T angle in diagnosing left ventricular hypertrophy. J Electrocardiol 2012;45:154-160. [63] Man S, De Winter PV, Maan AC, Thijssen J, Borleffs CJ, van Meerwijk WP, et al. Predictive power of T-wave alternans and of ventricular gradient hysteresis for the occurrence of ventricular arrhythmias in primary prevention cardioverter-defibrillator patients. J Electrocardiol 2011;44:453-459. [64] Burattini L, Man S, Burattini R, and Swenne CA. Comparison of Standard versus Orthogonal ECG Leads for T-Wave Alternans Identification. Ann Noninvasive Electrocardiol 2012;17:130-140.

ACCEPTED MANUSCRIPT [65] Borleffs CJ, Scherptong RW, Man SC, van Welsenes GH, Bax JJ, van EL, et al. Predicting ventricular arrhythmias in patients with ischemic heart disease: clinical application of the ECGderived QRS-T angle. Circ Arrhythm Electrophysiol 2009;2:548-554.

RI P

T

[66] de Bie MK, Koopman MG, Gaasbeek A, Dekker FW, Maan AC, Swenne CA, et al. Incremental prognostic value of an abnormal baseline spatial QRS-T angle in chronic dialysis patients. Europace 2013;15:290-296.

SC

[67] Kamphuis VP, Haeck ML, Wagner GS, Maan AC, Maynard C, Delgado V, et al. Electrocardiographic detection of right ventricular pressure overload in patients with suspected pulmonary hypertension. J Electrocardiol 2014;47:175-182.

MA NU

[68] Henkens IR, Mouchaers KT, Vliegen HW, van der Laarse WJ, Swenne CA, Maan AC, et al. Early changes in rat hearts with developing pulmonary arterial hypertension can be detected with three-dimensional electrocardiography. Am J Physiol Heart Circ Physiol 2007;293:H1300H1307. [69] Hillebrand S, de MR, Christen T, Maan AC, Jukema JW, Lamb HJ, et al. Body fat, especially visceral fat, is associated with electrocardiographic measures of sympathetic activation. Obesity (Silver Spring) 2014;22:1553-1559.

ED

[70] Boutkan,J. Vectorcardiography: physical bases and clinical practice. 1st ed. Eindhoven: Centrex Pub. Co; 1965

PT

[71] Malmivuo J and Plonsey R. Distortion Factors in the ECG. In: Malmivuo J and Plonsey R, editors. Bioelectromagnetism. Principles and Applications of Bioelectric and Biomagnetic Fields, New York, Oxford: Oxford University Press; 1995, p. 313-319

AC

CE

[72] FRANK E. The image surface of a homogeneous torso. Am Heart J 1954;47:757-768.

ACCEPTED MANUSCRIPT Figure legends Figure 1. Body surface isopotential map; modified[18] from Waller[17]. This isopotential and current flow

T

lines as reconstructed by Waller demonstrate the dipolar nature of the cardiac electrical activity. A: point

RI P

of the most positive potential, the localization of which is assumed to correspond to the intracavitary position of the apex of the heart; B: point of the most negative potential, corresponding to the base of the

SC

heart; a: isopotential lines, positive potentials; b: isopotential lines, negative potentials; c: current flow

MA NU

lines.

Figure 2. Vector representation of the electrical activity of the heart at a given moment; Figure 22 from Einthoven[6], with permission. Actually, the vector magnitude equals the segment pq (we marked this in Einthoven’s original figure by a blue line) on the shaft of the arrow, and not the length of the entire arrow.

ED

Vector projection on the three sides of the equilateral “Einthoven” triangle yield the amplitudes p1q1, p2q2, p3q3 in leads I, II, and III, respectively.

PT

Figure 3. Burger’s scalene triangle; Figure 37 from Boutkan [70], with permission. Analogous to Einthoven,

CE

Burger made this triangle projection representation to illustrate the consequences of the lead vector concept for Einthoven’s theory. In Burger’s triangle, the lengths of the sides (inner triangle with closed

AC

corners) represent the lead strength (sensitivity), and the direction of the sides represent the direction in which the leads “see” the heart vector (H). Comparing the Burger and Einthoven triangles reveals two basic differences. First, the directions in which the leads see the heart vector are not neatly 60° different: lead I assumes a slight upward direction, and leads II and III make a very sharp angle (and appear, compared to Einthoven, to be more in the same direction). Secondly, there is a remarkable difference in the sensitivities of the leads, lead I being least sensitive, lead III having the highest sensitivity. The difference in the lead vector directions (the directions of the sides of the inner triangle) cause a different projection of the heart vector on the inner triangle (compare with Einthoven’s projection in Figure 2). Moreover, after projection, the resulting vectors have to be multiplied by the strength of the leads (the relative length of the sides of the inner triangle). Arbitrarily taking, in this graphical representation, the strength of lead vector I as reference, the voltage in this lead, I’, is straightforwardly found by the

ACCEPTED MANUSCRIPT projection of the heart vector H on that lead vector. Leads II and III have larger strengths (longer sides of the inner triangle). Hence, the projections of the heart vector on the lead vectors II and III (II’ and III’, respectively) have to be amplified by the relative strength of these lead vectors with respect to the

RI P

T

strength of lead vector I to find the voltages in these leads (II and III, respectively). These multiplications have graphically been realized in this Figure by the help of an extra outer triangle with sides parallel to the

SC

inner triangle at appropriate distances.

Figure 4. Lead vectors of the 12-lead ECG; modified after Figure 18.1 from Bioelectromagnetism[71] , with

MA NU

permission. Components of the extremity lead vectors I, II, III, aVR, aVL and aVF that are generally assumed to be in the frontal plane are actually also in the transverse and sagittal planes. Components of the precordial lead vectors V1-V6 that are generally assumed to be in the transverse plane are actually also in the frontal and sagittal planes. Moreover, striking differences in lead strengths are seen. Also, the

ED

directions of the lead vectors differ sometimes dramatically from the idealized directions as usually depicted in text books (in the frontal plane aVL, I, II, aVF, III and aVR at 2, 3, 5, 6, 7 and 10 o’clock, and in

PT

the transverse plane V6-V1 at 2-7 o’clock, respectively).

CE

Figure 5. Scalar and Lissajous representations of the vectorcardiogram (VCG). Left panels, from top to bottom: scalar representations of the X, Y, and Z leads and of the vector magnitude (VM). Upper middle

AC

and right panels: 2D vector loops in the frontal, transverse and sagittal planes. Lower middle panel: 3D vector loop. Calibration: 0.5 mV/division. Colors mark the intervals between characteristic time instants in the ECG. Light red: onset QRS – instant of maximal QRS vector; dark red: instant of maximal QRS vector – end of QRS; light green: end of QRS – instant of maximal T vector; dark green: instant of maximal T vector – end of T; blue: ECG signal outside the QRS-T complex. The yellow arrow in the 3D vector loop indicates the maximal QRS vector. The magnitude of the maximal QRS vector is indicated in the 2D representations as a horizontal yellow line. The light blue-dotted lines in the 2D vector loops mark the lead-dependent maxima.

Figure 6. VCG lead system according to Frank[30], with permission. The 7 electrodes and the resistor network constitute a “corrected” lead system, i.e., lead vectors with equal strengths and pointing in the

ACCEPTED MANUSCRIPT direction of the main axes of the body. The directions of the x, y and z-axes correspond to what later became the American Heart Association standard[39]. Each electrode contributes to each of the X, Y and Z leads, except for the “head” electrode that is exclusively used for the derivation of the Y lead. H= head; F

RI P

T

= left foot; A, C, E, I and M are a selection of the electrodes that Frank used for an earlier study[72] where he placed electrodes all around the thorax in an anatomical transverse plane, designated in alphabetical

SC

order (A-P) and separated by 22.5° angles.

Figure 7. Pictorial summary of definitions of vectorcardiographic axes, planes and angles according to the

MA NU

AHA standardization recommendations [39]; Figure A1 from Man[60], with permission. X = vectorcardiographic x-axis (the arrow denotes the positive x direction); Y = y-axis; Z = z-axis; F = frontal plane; T = transverse plane; S = sagittal plane. As an example, an arbitrarily chosen instantaneous heart vector, H (red) is projected on the transverse plane (blue dotted line). The angle between the x-axis and

ED

this projection is the azimuth, A, of that heart vector (positive when turning in the direction of the positive z-axis). The angle between this projection and the heart vector is the elevation, E, of that heart vector

PT

(positive when turning in the direction of the positive y-axis). Hence, the heart vector as drawn in this

CE

Figure has a positive azimuth, a positive elevation, and positive amplitudes in the x-, y-, z directions.

Figure 8. Illustration of the synthesis of a VCG from a 12-lead ECG by the Kors matrix. The originally

AC

measured electrocardiogram, mathematically synthesized vectorcardiogram and Kors weighting factors for this synthesis are represented by matrices ECG, VCG and M, respectively. The 12-lead electrocardiogram is represented by matrix ECG that has 8 columns; each of these columns a-i represents the ECG signal in one of the 8 independent leads I, II, V1-V6, respectively. The matrix has m rows, where m indicates the number of samples in the signal. Typically, in a routine clinical 10-s ECG sampled at 500 Hz, m = 5000. The synthesized vectorcardiogram is represented by matrix VCG that has 3 columns; each of these columns j-l represents one of the 3 orthonormal VCG leads X, Y and Z, respectively. The number of samples equals that of the electrocardiogram. The Kors matrix consists of 3 columns in each of which 8 multiplication factors are found to compute a weighted sum of the samples of the 8 ECG leads as follows (example for sample i):

ACCEPTED MANUSCRIPT Lead X :    ji       .38   a i    –  .07   bi    –   .13  ci   .54  ii

T

Lead Y :   ki     –  .07     a i    .93   bi    .06   ci     .13  ii

SC

RI P

Lead Z :    li       .11   a i    –   .23   bi   –   .43  ci   .31  ii

Figure 9. Example of a synthesized VCG (lower panel), computed from a standard 12-lead ECG (upper

MA NU

panel) by multiplication of the 8 independent ECG leads I, II, V1-V6 by the Kors matrix (see Figure 8 for the calculations). For an initial orientation it is useful to inspect the Kors matrix in Figure 8 to find the most important ECG leads that contribute to a given VCG lead. ECG leads V6 and I have major contributions to VCG lead X, with weighting factors of .54 and .38, respectively. ECG lead II has a major contribution to

ED

VCG lead Y, with a weighting factor of .93. ECG leads V1 and V6 have major contributions to VCG lead Z, with weighting factors of .43 and .31, respectively (note that the ECG lead V1 weighting has a minus sign

PT

which means that the lead must be inverted). Visual comparison of lead X with leads V6 and I, of lead Y

CE

with lead II, of lead Z with inverted lead V1 and with lead V6 shows clear resemblance, however it has to be realized that each of the ECG leads contributes to each of the VCG leads and that the correctness

AC

(orthonormality) of the synthesized VCG actually requires the contributions by all ECG leads, also the smaller ones. Strikingly, in this example, the peak-peak amplitudes in the X, Y and Z leads have comparable magnitudes, while in the ECG the differences are large: e.g., compare lead I to leads V3/V4; these differences are caused by the strongly different lead strengths in the electrocardiogram (see Figure 4).

Figure 10. Illustration of the spatial QRS-T angle, normal subject (female, 20 years). The Figure shows the QRS- and the T loops (red and green, respectively; the T loop in this example is relatively narrow, as is often seen in normal subjects). The QRS- and T axes are the spatial orientations (azimuth and elevation) of the QRS- and T integrals (vectorial additions of the QRS- and T areas in the X, Y and Z leads, respectively). The three depicted views (from left to right: right-anterolateral, anterior and left-anterolateral) show

ACCEPTED MANUSCRIPT clearly the influence of the projection: the “genuine” QRS-T angle in this subject, 81°, is best seen in the right-anterolateral view while the frontal view shows a QRS-T angle of only 43°. This demonstrates that

T

the QRS-T angle should always be measured in 3D space and not in the frontal plane.

RI P

Figure 11. Illustration of the computation of the ventricular gradient (VG); Figure 2 from Ter Haar et al.[61], with permission. ECG of a patient with acute ischemia as a consequence of balloon occlusion of a

SC

coronary artery segment during elective percutaneous transluminal coronary angioplasty. Panel A: synthesized VCG (X, Y and Z leads) and vector magnitude. Vertical time markers indicate onset

MA NU

QRS, J point, J + 60-ms instant (red) and end of the T wave. The x, y, and z components of the VG vector are computed as the areas under the curve of the QRS-T complex; positive amplitudes (petrol) contribute positively and negative areas (purple) contribute negatively to the area. In this example, the net QRS-T areas of the X and the Y leads are obviously positive, while the net QRS-T area of the Z lead is negative.

ED

The latter is due to the ST depression and the negative T wave.

Panel B: composition of the VG vector (vector sizes in mV∙ms). The magnitudes of the x, y and z

PT

components, VGx, VGy and VGz, equal the net QRS-T areas as illustrated in panel A. VGx and VGy are

CE

positive and point, hence, in the positive directions of the x- and the y axes. VGz is negative and points,

VG vector.

AC

hence, in the negative direction of the z-axis. Vectorial summation of VGx, VGy and VGz yields the resultant

Figure 1

AC

CE

PT

ED

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

Figure 2

CE AC

Figure 3

PT

ED

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

Figure 4

ED

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

PT

Figure 5

AC

Figure 6

CE

PT

ED

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

Figure 7

CE

PT

ED

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

PT

ED

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

Figure 8

ED

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

PT

Figure 9 lower

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

Figure 9 upper

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

Figure 10

AC

Figure 11

CE

PT

ED

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT Sumche Man, MD, Arie C. Maan, PhD, Martin J. Schalij, MD, Cees A. Swenne, PhD:

RI P

T

Vectorcardiographic diagnostic & prognostic information derived from the 12‐lead electrocardiogram: historical review and clinical perspective

CE

PT

ED

MA NU

Historical overview of the development of electrocardiography and vectorcardiography Origin and evolution of the heart vector and lead vector concepts Decline and revival of vectorcardiography Vectorcardiography as an extension of the standard 12-lead electrocardiogram Current and future clinical and research applications of the vectorcardiogram

AC

-

SC

Highlights: