Introduction

Cardiac electrical activity can be recorded on the body surface by devices such as electrocardiographs and vectorcardiographs1. In 1888, Waller2,3 defined the concept of the electric vector and conducted the first experimental studies on his dog Jimmie. In 1902, Einthoven4 enhanced this method for humans, modernizing its application.

The American engineer Ernest Frank (1956)5 created the vectorcardiographic derivation system using an orthogonal method of corrected derivations. At this time, veterinary medicine began its electrocardiography (ECG) studies. Several methodologies were proposed to study the electrical activity of the canine heart through the electrocardiogram and vectorcardiogram. However, the practical positioning of the electrodes was continuously adapted from bipedal to quadrupedal animals6,7,8,9,10,11,12,13,14.

John Hill6 mentions that Lannek defined the studies on canine ECG in 1949, including the precordial methodology7. Later, in 1964, Detweiler and Patterson proposed a standardization of electrocardiographic methods7,8 by performing the electrocardiogram examination in the right lateral decubitus position. A consensus was reached that the electrodes should be positioned on the front and back limbs, as proposed for humans4,5,6,7,8.

Takahashi, in 1964, proposed the system of twelve precordial leads, six in the left hemithorax and six in the right9. In 1965, Detweiler and Patterson7,8 modified the methodology described by Takahashi9 by reducing the number of electrodes. Hill6, in 1968, in research with electrodes on the limbs of dogs, adopted different positions to evaluate how the modification in the position of the paws of the forelimbs could affect the electrical axis of the QRS complex.

Nunes, Moffa, and Iwasaki, in 1990, carried out experimental studies with vectorcardiography in dogs11. However, concerning the frontal plane, with electrodes on the limbs, they followed what is recommended by Lannek6,7,8,9,10, Takahashi9 and Hill6,7,8,9,10. Kraus12 in 2002, using the 12-lead ECG, proposed combining the methods recommended by human medicine, that is, maintaining the use of electrodes on the limbs but modifying the precordial methodology, applying the same positioning and quantity of electrodes proposed by Wilson7 for the thorax of dogs. The researcher considered the different shapes of canine thoraxes compared to the human thorax12. The hypothesis was that the proposed electrode arrangement would be able to distinctly assess the relationship between the right ventricle, interventricular septum, and left ventricle, following what is determined in human medicine12.

In a recent study published in 2019, Roberto Santilli et al13 aimed to evaluate atrial and ventricular activation using a 12-lead ECG across different canine chest shapes. They modified the transverse plane recommended by Kraus12, placing electrode V1 in the first intercostal space on the right, unlike Kraus who positioned it in the fifth intercostal space on the same side. The latter method is currently the conventional practice for canine electrocardiography. Santilli’s13 objective with this new positioning was, to better represent atrial and ventricular activity in the most varied chest shapes such as brachiomorphic, mesomorphic, and dolichomorphic16,17,48,49.

Thus, considering the current literature, it is essential to revaluate the classical methods of ECG evaluation in dogs since the traditional arrangement of electrodes suggested to date does not consistently define the correct spatial position of the canine heart.

Therefore, the present study aims to evaluate a new distribution of electrode positioning in the canine thorax through electrovectorcardiography and compare it with the conventional method13.

Results

These results represent a preliminary methodological step with translational potential. However, further validation in dogs with heart disease is required to clarify their diagnostic and prognostic value. We evaluated 90 dogs (purebred and mixed breed): 50 females (55.6%), with a mean weight of 17.5 ± 10.2 kg, a mean height of 0.40 ± 0.11 m, and a mean age of 2.6 ± 1.36 years. The study included a wide variety of defined breeds. This variety enabled a more comprehensive assessment of morphofunctional variations that may impact electrocardiographic and hemodynamic parameters. Dogs were grouped by thorax width: 21 brachycephalic (23.3%), 60 mesocephalic (66.7%), and 9 dolichocephalic (10%) (Supplementary Table S2). The breeds included were Mixed Breed (n = 27), English Cocker Spaniel (n = 9), Bulldog (n = 9), Greyhound (n = 8), Malinois Shepherd (n = 6), American Bully (n = 5), Golden Retriever (n = 5), Border Collie (n = 4), Yorkshire (n = 4), Shih Tzu (n = 4), Pitt Bull (n = 3), Doberman (n = 1), Poodle (n = 1), Maltese (n = 1), Dachshund (n = 1), German Shepherd (n = 1), and Rottweiler (n = 1).

The analysis of variance (ANOVA) test demonstrated that, despite variations in thoracic morphology, there was no significant difference in the ƩQRS (mV) values among the three groups, regardless of the positioning method used (new and conventional). Therefore, we decided to combine the data from all ninety dogs into a single group to evaluate the two distinct methods of electrode positioning.

Comparison of electrocardiographic findings in the frontal plane between the new and conventional methods

The P wave in DI was 0.15 mV ± 0.07 mV (reference range 0.10 to 0.20 mV) and in AVL was 0.10 mV ± 0.06 mV (reference range 0.06 mV to 0.10 mV) (p < 0.001), being larger compared to the conventional Santilli method that showed DI of 0.07 mV ± 0.06 mV (reference range 0.0 mV to 0.10 mV) and AVL of 0.06 mV ± 0.07 mV (reference range 0.0 mV to 10 mV). In leads DI, DII, DIII, and AVF, the conventional method revealed the presence of abnormal Q waves, whereas the new method did not detect any abnormal phenomenon in these same leads (p < 0.001).

The new method showed a greater amplitude of the R waves in lead DI 0.82 ± 0.33 mV (reference range: 0.06 mV to 1.02 mV), DII 1.60 mV ± 0.48 mV (reference range: 1.23 mV to 1.90 mV), DIII 0.90 mV ± 0.36 mV (reference range: 0.70 mV to 1.12 mV), and AVF 1.19 mV ± 0.42 mV (reference range: 0.9 mV to 1.4 mV) (Table 1 and Fig. 1A, B).

Table 1 Frontal plane electrocardiographic parameters.
Full size table
Fig. 1
figure 1

Lead electrocardiogram. (A) Conventional Method; (B) New Method.

Full size image

The S waves leads obtained by the new arrangement were negative and were present in all ninety dogs showing in the DI mean values of − 0.08 mV ± − 0.13 mV (reference range: − 0.10 mV to 0.0 mV), DII − 0.33 mV ± 0.25 mV (reference range: − 0.50 mV to 0.20 mV), DIII − 0.35 mV ± − 0.23 mV (reference range − 0.50 mV to 0.20 mV) and AVF − 0.33 mV ± 0.24 mV (reference range − 0.50 mV to 0.20 mV). Furthermore, the final portion of the QRS, characterized by the J point, is visually present in the new method. In contrast, in the conventional method, it is not clearly and evidently detected (Table 1).

The DI ƩQRS values comparing the new method to the conventional one showed a mean amplitude 0.74 mV ± 0.37 mV (reference range: 0.46 mV to 0.10 mV) vs 0.25 mV ± 0.24 mV (reference range: 0.10 mV to 0.40 mV), respectively (p < 0.001). Additionally, the ƩQRS was significantly higher in millivolts in the new methodology (p < 0.001) (Table 1).

The T waves in the frontal plane presented a uniquely positive aspect in the DI with mean values of 0.15 mV ± − 0.18 mV (reference range: 0.10 mV to 0.30 mV), DII 0.49 ± 0.28 mV (reference range: 0.30 mV to 0.60 mV), DIII 0.34 mV ± 0.20 mV (reference range: 0.20 mV to 0.49 mV), and AVF 0.42 mV ± − 0.24 mV (reference range 0.26 mV to 0.56 mV) leads (p < 0.0001), in comparison to the conventional arrangement of electrodes that shows different millivolt values, in addition to different polarities such as positive, negative and biphasic (Table 1).

Comparison of electrocardiographic findings in the transverse plane between the new method and the conventional one

In the transverse plane, the P wave recorded from leads V1 was − 0.07 mV ± − 0.06 mV (reference range: -0.1 mV to − 0.06 mV) and V2 was − 0.02 mV ± − 0.08 mV (reference range: − 0.1 mV to 0.05 mV) showing a more negative amplitude using the new method (p < 0.001), while in the arrangement from V3 the results was 0.03 mV ± − 0.07 mV (reference range: − 0.04 mV to 0.08 mV), V4 was 0.07 mV ± 0.04 mV (reference range: 0.06 mV to 0.10 mV), V5 was 0.09 mV ± − 0.06 mV (reference range: 0.08 mV to 0.10 mV), V6 was 0.09 mV ± 0.03 mV (reference range: 0.08 mV to 0.10 mV), and the P wave appeared positive (Table 2).

Table 2 Transverse plane electrocardiographic parameters.
Full size table

The new method recorded R waves in V1 of 0.16 ± 0.13 mV (reference range: 0.08 mV to 0.20 mV) and in V2 of 0.28 mV ± 0.19 mV (reference range: 0.12 mV to 0.40 mV) when compared to the conventional method that showed in V1 mean values of 0.56 mV ± 0.30 mV (reference range: 0.30 mV to 0.72 mV) and V2 of 0.99 mV ± 0.50 mV (reference range: 0.70 mV to 1.30 mV). The leads V3 showed values of 0.50 mV ± 0.25 mV (reference range: 0.30 mV to 0.70 mV), V4 of 0.74 mV ± 0.31 mV (reference range 0.51 mV to 0.90 mV), V5 of 0.95 mV ± 0.32 mV (reference range 0.70 mV to 1.10 mV) and V6 of 0.98 mV ± 0.34 mV (reference range: 0.75 mV to 1.20 mV), thus the R waves increase their amplitude progressively. Lead V6 with the new electrode arrangement demonstrates a mean QRS of 0.81 mV ± 0.39 mV (reference range: 0.60 mV to 1.00 mV) (p < 0.001), compared to the conventional method (0.22 mV ± 0.43 mV; reference range: 0.0 mV to 0.40 mV) (p < 0.001). The J point is present and more clearly visible in V1 to V6 using the new method (p < 0.001) (Fig. 1A, B). The T waves from V1 0.10 mV ± 0.19 mV (reference range: − 0.02 mV to 0.20 mV), V2 0.18 mV ± 0.23 mV (reference range: 0.09 mV to 0.30 mV), V3 0.24 mV ± 0.21 mV (reference range: 0.12 mV to 0.32 mV), V4 0.29 mV ± 0.21 mV (reference range: 0.14 mV to 0.40 mV), V5 0.35 mV ± 0.63 mV (reference range: 0.14 mV to 0.40 mV), and V6 0.25 mV ± 0.18 mV (reference range: 0.12 mV to 0.34 mV) have a positive aspect, while by the conventional arrangement, they decrease from V3 0.22 mV ± 0.24 mV (reference range: 0.10 mV to 0.40 mV), V4 0.13 mV ± 0.38 mV (reference range: 0.0 mV to 0.20 mV), V5 0.0 mV ± 0.23 mV (reference range: − 0.10 mV to 0.10 mV) to V6 − 0.07 mV ± 0.18 mV (reference range: − 0.20 mV to 0.0 mV).

Vectorcardiographic findings in the frontal plane between the new method and the conventional one.

The vectorcardiographic loop of the P wave, using the new method, presents an average angulation of 3.5° ± 37.3° (reference range − 7.5° to 24°), whereas using the conventional method, there is a wide variation (41.9° ± 49.2°; reference range: 39° to 64°).

The QRS loop axis showed a 35% variation between the new arrangement and the conventional one (mean of 30.1° ± 24.2°; reference range: 23° to 42°) vs 22.4° ± 72.1° (reference range: 28° to 52°), respectively.

The T loop showed a difference of 44% in relation to the conventional method (53.6° ± 46.5° (reference range: 45° to 71°) vs − 95.2° ± 110.1° (reference range: − 159°, − 119°) (p < 0.001) (Fig. 2 AB).

Fig. 2
figure 2

Frank VCG system. (A) Conventional method. (B) New electrode arrangement method.

Full size image

Comparison of vectorcardiographic findings in the transverse plane between the new method and the conventional one

The degree of orientation of the P loop showed a difference of 31% in the comparison between the positioning methods, the mean is − 29.6° ± 36.8° (reference range − 44° to − 19°) for the new arrangement and 42.9° ± 51.0° (reference range: − 3° to 95°) for the conventional arrangement. Regarding the QRS loop, in the comparison between the methods, with the mean being 69.4° ± 56.7° (reference range: 9.5° to 113°) for the conventional arrangement and − 20.4° ± 26.8° (reference range: − 34° to − 9°) for the new arrangement.

As for the T loop, this difference amounted to 62%, with a mean of 45.1° ± 108.5° (reference range: − 4° to 80°) in the new arrangement and a wide variation with a mean of 121.2° ± 34.6° (reference range: 113° to 138°) in the conventional arrangement. In the sagittal plane, the P, QRS, and T complexes were shown to agree with the frontal and transverse planes in the new arrangement. In contrast, in the conventional arrangement, there is discordance mainly in the T loop, which is negative (− 23.6° ± 26°) (Fig. 2 AB).

Discussion

There are several recommended methodologies for arranging electrodes in dogs to this day. Still, all of them use the adapted form of human cardiology, that is, for bipeds. The new method is no longer an adaptation but rather a new methodology that respects the principles of Waller, Einthoven, Frank, and all the anatomy of the planes and axes, as well as all canine cardiac electrophysiology. However, several authors8,9,10,11,14,15 consider the wave shapes according to the conventional methodology to be normal, such as P waves of various morphologies, Q with deep voltages, QRS complexes alternating in voltages, ST segments and T waves with different shapes, positive, isoelectric and negative, alternating in the electrocardiographic tracing of the same dog6,8,9,11,12,13,14,15. Based on healthy dog hearts, the authors19,20,21 consider the mean electrical axis of the QRS in the frontal plane between 40° and 100°, but Catcott22 discusses in his study that the axis varies from 0 to 104°. While the conventional method demonstrated dispersion of the QRS axis from 0 to 104°, the new method revealed an angle of 5.9° to 54.3° with a mean concentrated at 31.1°, reflecting greater consistency and electrocardiographic stability.

This deviation is the subject of studies nowadays21,23, once that canine thoracic conformations (brachiomorphic, mesomorphic, and dolichomorphic) influence the mean electrical axis of the QRS and the P and T waves by the conventional methodology. By respecting the canine anatomy, this new arrangement proved to be a physiologically assertive method for electrovectorcardiographic analysis.

Einthoven’s triangle should project the vectors vertically. However, the conventional method processes the results distorted because the electrodes are on the limbs. This inverts the capture of electrical activity, presenting low visualization through the DI derivation.

The vectorcardiographic loop of the P wave in the frontal plane using the new method showed a mean angle of 3.5° ± 37.3° degrees, which proves a better visualization of the P wave using the new methodology compared to the conventional one (41.9° ± 49.2°). According to Borgarelli, Bussadori, and Santilli24, its electrical axis should be from − 18 to 90°. Thus, it is determined that in the DI lead, the P wave can be positive, diphasic, or even isodiphasic, and in the AVR and AVL leads, it assumes negative polarity. This finding would make interpretation difficult since for the rhythm to be considered sinus rhythm, the P wave must be positive in DI, DII, and AVF25,26.

In the new method, the derivation that evaluates the left lateral as DI was twice as high in millivolts as the conventional method. The DII and AVF derivations showed positive P waves (p < 0.0001), respecting the electrophysiology and electrovectorcardiography. In the AVL derivation, the R wave increased amplitude compared to that proposed by the conventional method (0.35 mV ± 0.23 mV vs 0.28 mV ± 0.21 mV).

According to Santilli14, the first vector of the interventricular septum is oriented from inferior to superior and from left to right in the frontal plane. By the conventional method, the mean electrical axis was shown to be between + 40° and + 100° in 80% of dogs. Thus, the electrical activity seen by leads DI, DII, DIII, and AVF registers the negative deflection as a Q wave.

In our study, according to the conventional ECG and VCG methodology, the QRS waves/loops begin upwards and to the right, representing the inscription of Q waves on the ECG. Subsequently, the loop passes through the positive portion of the DI, DII, DIII, and AVF leads, demonstrating R waves, and ends in the basal portion, forming the QR complex. The beginning of repolarization also has a direction to the right and upwards, configuring negative T waves in the inferior leads.

According to Samesima et al.29 and Pastore et al.30,34, the presence of Q waves, considered pathological, in the inferior wall represents an inactive electrical area/fibrosis, evidencing a zone of necrosis due to electrolytic and metabolic imbalance27,28. Comparatively, using the new method, the DI derivation presents a means by which the appearance of Q waves in the inferior wall was undetected. We can conclude that the more reliable results presented by the new arrangement are because the dog’s heart is now assumed to be in the center of Einthoven’s triangle and that through the VCG loop by the Frank system, septal activation occurs parallel to AVF, DII, and DIII, not recording Q waves.

On the studies of Samesima29 and Shepherd31,32,33, S waves are recorded in DI, DII, DIII, and AVF in the frontal plane, confirmed in canine electrocardiology by the new arrangement. The S waves were also present in most dogs in leads DII, DIII, AVF, V6, and AVR, compatible with end-conduction delay (ECD) diagnosis.

The ECD occurs when the right ventricle has slower than normal depolarization but without the right bundle branch block, and no forces are opposed to those of the left ventricle. The appearance of these waves on the ECG occurs due to divisional blocks of the right bundle branch or the existence of a region poorer in Purkinje fibers in the right ventricle outflow tract1,27,31,32,33,34,35. In this situation, the S of DII is greater than that of DIII because the terminal activation forces oppose this derivation. In addition to being under five years of age, these dogs are all healthy according to the echocardiogram, and this manifestation may be a variant of normality or a benign phenomenon.

As Santilli14,15 points out, the first vector has an orientation in AVR, and inscribing R waves becomes confusing because the same author emphasizes that septal activity occurs from inferior to superior and from left to right. Due to the generation of R waves in DII, DIII, and AVF, the same vector seen in AVR and AVL leads captures the electrical activity of the deep S wave. Due to the adaptive form of human medicine, how the electrodes are arranged in the conventional methodology indicates a vector inversion, with records of negative P waves and positive R waves, followed by S waves, but with a positive T wave discordant with P and QRS.

It is also possible to observe the T waves present in the DI, DII, DIII, and AVF leads in a negative way using the conventional method. While using the new arrangement, the ventricular repolarization represented by the T wave was positive in the DI, DII, DIII, and AVF leads in the ninety dogs. Although Santilli et al.13,14,15 consider different polarities in the T wave in dogs, in their most recent works, these authors indicate that how the heart is inside the thorax for the different thoracic morphologies may be the cause of this wide variation in polarity. Thus, when obtained by the conventional method, there is no consensus.

These results were not reproduced by the new positioning. A reliable analysis of cardiac electrical activation showed that it was not influenced by the morphological difference of the thorax.

The studies carried out by Antzelevitch et al.37,38,39 demonstrates in dog hearts that there is a sequential repolarization process that must be respected. Thus, ventricular repolarization can be understood as the sum of all action potentials of ventricular myocytes, repolarizing three layers, first the epicardium, passing through the endocardium and, finally, the M cells, which characterizes electrophysiologically the positive T wave and its asymmetrical appearance, thus being considered normal1,26,27,37,38. In an ischemic event, the T wave would configure a negative format due to the alteration of this entire sequence37.

According to Antzelevitch, in both humans and dogs, there is a normal time between the layers of the myocardium considered spatial dispersion of repolarization. This formation and generation of the transmural, transseptal, and apicobasal dispersion of repolarization creates voltage gradients responsible for producing the J point and the T wave of the electrocardiogram in a physiological manner37,38,39. Antzelevitch’s finding is opposite to what is considered normal by the conventional veterinary medicine method.

In the new method, the T waves are positively inscribed in all ninety animals and the different thoracic conformations, maintaining the pattern of positivity demonstrated by Antzelevitch37.

Using the new method, the electrodes were arranged horizontally, allowing rational and electrophysiological characterization of the orientation and direction of the electrical activity of the atriums and ventricles and the repolarization of the ventricles.

The VCG loop of the P wave constitutes the smallest vectorcardiographic loop, and its initial forces in the dog correspond to the right atrium, having an anterior, inferior (ventral), and slightly leftward (caudal) orientation in the frontal and transverse planes with a small initial anterior part and a late posterior component.

The QRS VCG loop is the composition of all instantaneous vectors of ventricular activation. In the frontal plane, septal activation always occurs in the upper and right portions, starting on the right side (in relation to the RV) and continuing to the left and downwards (ventral). In the transverse plane, septal activation always occurs in the anterior portion, starting with a vector to the right in relation to the RV. Subsequently, the electrical activity is directed to the left (left caudal). This is in accordance with the anatomical notion of the canine heart. In the horizontal plane, the left ventricle, with greater mass and more significant deflection on the ECG, is located posterior to the right ventricle. In addition, the VCG loop has a greater magnitude, ovoid or rounded appearance, and counterclockwise rotation. Activation of the free walls remains on the left, with a clear predominance of the location of the loop on the posterior side.

According to the new methodology, the T wave loop has an elliptical or elongated shape, with its efferent branch being much slower than the afferent branch. Its maximum vector is oriented inferiorly, to the left and forward, with the loop rotating in line with that of the QRS, i.e., counterclockwise in the transverse plane, clockwise in the sagittal plane, and variable in the frontal plane.

The new transverse plane method is clearer and easier to understand. It also shows that the R waves progress and grow while the S waves decrease from V1 to V6.

Another finding with the new arrangement is that despite the different thoracic morphologies, the electrocardiographic pattern was maintained, where ƩQRS showed to be, in addition to being positive, with greater amplitude where V5 and V6 presented better clarity in seeing the left side of the LV.

When electrodes are arranged using the conventional method, the electrical phenomenon is observed perpendicular to the electrical path of activation. Thus, the position of the electrodes using the conventional method remains vertical, and different results are obtained when applied to different chest morphologies.

Due to the new arrangement, following the electrical activation in the craniocaudal direction (right/left), the thorax’s conformation does not influence the results, regardless of whether it is brachiocephalic, mesocephalic, or dolichocephalic. With the new method, it was possible to understand the spatial orientation of the vectors in relation to the anatomy of the canine heart through the new planes and axes, configuring orthogonality.

It was clear that the new methodology, based on the principles established by the developers of ECG/VCG, places the heart inside the thorax in its anatomical position. There is no doubt that by obtaining ECG and VCG tracings that are very similar to the electrical activation of humans, the comparison with the already established standards favors us and avoids questionable patterns of the electrical phenomenon. Through the new method, it was possible to create an ECG/VCG pattern that respected the electrophysiology of the canine heart and facilitated the vectorial knowledge of the process.

The proposed new arrangements discuss how ECG/VCG in dogs can provide additional information on left/right chamber voltages, left and right bundle branch blocks, divisional blocks, and identify electrically inactive areas. Considering that diseases and comorbidities that affect both humans and dogs have similar clinical and phenotypic presentations, this new electrovectorcardiographic arrangement may, in the future, contribute to early diagnosis and cardiological screening in cases similar to those in human medicine. Although the data from this study indicate a possible methodological advantage, it is restricted to validation in healthy animals. Therefore, further studies are necessary in populations with cardiac diseases to validate the clinical applicability of the new method and establish specific reference intervals.

Conclusion

The new electrode arrangement methodology, based on electrovectorcardiography, offers an innovative perspective to electrocardiographic interpretation in dogs, regardless of thoracic conformation, with diagnostic potential and greater assertiveness of possible cardiac alterations.

Materials and methods

An experimental, cross-sectional, prospective clinical study was conducted with ninety healthy dogs without anesthetic intervention. Transthoracic echocardiogram (TTE), electrocardiogram (ECG), and vectorcardiogram (VCG) were performed at the university veterinary hospitals or private kennels. All dogs were of domestic origin, and the person responsible for the dog or the kennel randomly took them for an elective clinical surgical consultation.

First, we collected demographic data such as weight, gender, age, clinical evaluation, and, subsequently, the morphometry of each dog in a stationary position (standing), measuring the width and depth of the thorax using an anthropometric ruler with the subsequent classification of the morphology of the thoracic conformation.

The dogs’ heights were measured from the beginning of the forelimb (ground) to the withers region (proximal to the scapula at the height of the scapular cartilage) to index the weight associated with height on the echocardiogram to obtain the myocardial mass index. We measured the thoracic index (thoracic width X 100 / thoracic depth). Dogs with thoracic indexes of 90 to 100, 60 to 89, and 50 to 59 were classified into the brachiomorphic, mesomorphic, and dolichomorphic groups, respectively16,17,48,49.

The project was approved by the Ethics Committee on the Use of Animals (CEUA) of the Hospital das Clínicas da Faculdade de Medicina da USP (CEP_HCFMUSP; nº1540/2021) and by the CEUA of the Pontifícia Universidade Católica de Minas Gerais (CEUA-PUC-Minas). The study follows the precepts of Law nº 11.794 and the standards issued by the National Council for the Control of Animal Experimentation (CONCEA). All methods were performed in accordance with relevant guidelines and regulations, and the study is reported in accordance with ARRIVE guidelines.

Inclusion Criteria: Dogs aged between 1 and 5 years, of varying weights, of various breeds, and of both sexes were included. Exclusion Criteria: Dogs that presented any abnormality detected and diagnosed in clinical, physical, and/or cardiovascular anamnesis examinations by TTE and ECG, including congenital and acquired alterations. The same veterinary cardiologist performed all examinations.

Transthoracic echocardiogram (TTE)

The TTE helped assess the absence of cardiac chamber dilation and congenital alterations. The dogs were in the right and left lateral decubitus to visualize the right and left parasternal acoustic windows36,37,38,40,41,42. The portable model used was General Electric GE, model Vivid E, with cardiology software and multifrequency transducers 2.5 to 7 MHz (Supplementary Table S1).

Electrocardiography (ECG)

The electrocardiograms were recorded using the 12-lead ECGPC software, with ten leads, two speeds (50 mm/s), and three gain settings (5, 10, and 50 mV). The VCG vectorcardiogram was generated from the software module connected to the ECGPC system (TEB) (version VS1 RV8), developed by TEB Tecnologia Eletrônica Brasileira, São Paulo, Brazil. Available at http://www.teb.com.br. The speed established for the examination was 50 mm/s, and the gain was 10 mV. Metal electrodes were placed on the animal’s skin with 70% alcohol to capture electrical activity. The examinations were performed with the dog in the right lateral decubitus position. We performed both exams (TTE and ECG) in two stages, the first of placing the electrodes according to the conventional method and, in the second stage, according to the new proposal.

Electric Axis: The electrical axis result was evaluated following the principles of trigonometry, which, through the values (leg, opposite leg, and hypotenuse), evaluates the amplitudes of the QRS complex of the DI and AVF leads, as used in human medicine27,28,50,51.

Wave amplitude: The QRS complex represents the depolarization of the ventricles. The typical QRS has three components: Q waves as the first negative deflection, R waves as the positive deflection, and S waves as the second negative deflection. Voltage analysis was performed using the algebraic sum of its amplitudes in millivolts (ƩQRS) in all electrocardiographic leads of the two planes18.

Conventional 12-lead electrocardiogram method (santilli)

Following the guidelines of Santilli13, the dogs were placed in right lateral decubitus, and the electrodes were arranged as follows for the frontal plane: yellow electrode on the left forelimb, red electrode on the right forelimb, green electrode on the left hind limb and black electrode on the right hind limb. For the horizontal plane, the electrodes were arranged so that electrode V1 was cranially in the 1st intercostal space on the right side of the thorax, and electrodes V2 to V6 were placed on the left hemithorax in the sixth intercostal space.

New method for positioning the 12-lead electrocardiogram

Einthoven’s triangle was arranged in the following manner in the frontal plane: red electrode (negative) dorsal to the scapula in the region of the scapular cartilage, yellow electrode (positive) caudally to the last rib (floating lumbar region), the other green electrode (positive) in the area of the xiphoid process and the black electrode (neutral) in the left forelimb in the middle third (Fig. 3). In the horizontal plane, the unipolar electrodes were placed from V1 in the first, V2 in the second, V3 in the third, V4 in the fourth, V5 in the fifth space, and V6 in the sixth left intercostal space, close to the costochondral junction, following a linear arrangement of the same (Fig. 3).

Fig. 3
figure 3

The new method of electrode arrangement in dogs is a 12-lead electrocardiogram. (A) Frontal plane. (B) Transverse plane.

Full size image

The dogs were placed in the same position as the conventional method (right lateral decubitus), and new topographic references were proposed for the frontal plane, such as superior dorsal, inferior ventral, right cranial, left caudal, and for the horizontal plane, anterior left and posterior right (Fig. 4).

Fig. 4
figure 4

The new electrode arrangement of the Frank system for dogs, along with ECG/VCG topography on the frontal and transverse planes.

Full size image

The variables studied were the P, Q, R, S, ƩQRS, J point, T, and U waves in their millivolts. For the analysis, we considered the six derivations of the frontal plane and six derivations of the horizontal plane. The quantitative analyses were performed by two examiners, veterinary cardiologists.

Vectorcardiogram

The VCG was obtained from a software module coupled to the ECGPC model (TEB). From the tracing through the 7-lead Frank system cable, transformations of the Frank leads occur, corresponding to the vectorcardiogram1,26,43,44,52. The module expresses the vectorcardiographic loops of P, QRS, and T in the three orthogonal planes X, Y, and Z decomposed into frontal, transverse, and right sagittal planes.

The software measured the cubic measurements of the P, QRS, and T loops and the vector projection in each P, QRS, and T plane. The angles and axes of the P, QRS, and T loops and the direction of vector rotation for each vectorcardiographic loop were also evaluated26,43,52,53.

The tests were performed with the dog in the right lateral decubitus position, lying down, and in two moments, the first consisting of placing the electrodes according to the conventional method and the second according to the new proposal for electrode placement1,52. The new arrangement creates loops that are related to the 12-lead electrocardiogram. The first vectorcardiographic loop to be activated represents the electrocardiogram’s P wave, a single wave of small amplitude, denoting the right and left atrial activity. In the second instance, another open loop of large proportion, circular, oval, or even triangular appearance forms the QRS complex of the electrocardiogram, in which it shows the ventricular systole1,26,43,52,53,54. The last loop created represents the T wave, which in the ECG has an asymmetric and rounded morphology, representing the ventricles repolarization.

The focus was on analyzing the QRS loop (clockwise/counterclockwise rotation, duration, orientation) and the J point (position, amplitude of the resulting vector, and angle formed between the end of the QRS loop and the beginning of the T wave). The resulting vector of the J point was obtained through the non-coincidence of the start of the QRS loop and its end. Additionally, the amplitude of this vector and its angle were measured1,26.

Orthogonal plane—frank system for conventional veterinary medicine

Santilli14,15, as well as Calvert55, and other authors56,57,58,59 describe that the electrodes should be placed as follows: the negative electrode X in the 5th right intercostal space, the positive electrode in the 6th left intercostal space, the negative electrode Y placed in the manubrium region, the positive electrode Y on the xiphoid cartilage, the positive electrode Z on the spinous process of the thoracic vertebra T7 and the negative Z electrode ventral region on the sternum.

Orthogonal plane—snew vetocardiogram method for veterinary medicine

The electrodes used followed the arrangement format recommended by Frank5 in the 5th intercostal space and other points of the body. The letters (I, E, C, A, M, H, and F) represent the electrodes arranged on the thorax. However, for the study proposal in question, for the canine species, these same electrodes were placed in a different arrangement following the thorax’s anatomy, following Frank’s principles5 (Fig. 4). Through the new proposed arrangement of the electrodes, the following shape is given for the dog’s thorax: the dog in the right lateral decubitus position, with the electrodes I 1°, E 3°, C 5°, A 7° in the left intercostal space at the level of the costochondral junction. The electrode H caudally to the last rib (floating, left lumbar region), electrode M in the right mesogastric region in the middle third of the abdomen, electrode F in the xiphoid process, and electrode N in the area of the middle third of the left forelimb (Fig. 4). This arrangement aimed to centralize the heart through several resistances forming the orthogonal planes (Fig. 5).

Fig. 5
figure 5

Planes and axes. (A) Planes and axes of the human body. (B) Adaptation of the planes and axes of the dog’s body. (C) and (D) New planes and axes utilizing electrovectorcardiography.

Full size image

Studied variables of the vectorcardiogram

The variables analyzed by the VCG included the orientation, angulation (degree), rotation, and spatial amplitude (mV) of the P loop, QRS T in the transverse, frontal, and right sagittal planes, as well as the angle formed between the J point and the beginning of the T wave, in the three planes. A cardiologist from the electrocardiography unit of HeartInstitute (InCor- HCFMUSP) analyzed the VCGs, and in case of doubt, another examiner analyzed the results.

Sample size

To calculate the sample size, we used the magnitude of the R wave in D2 as the primary outcome, with a mean value of 1.31 millivolts (mV) with variability of 1.08 mV (SD = 1.08 mV) observed in the literature60. A difference of at least 0.5 mV is expected in the R wave in D2 with the new electrode positioning method, with a statistical power of 85% and 95% confidence. The sample required for the study is eighty-four animals, considering a two-tailed test for paired samples.

Statistical analysis

The analyses were performed using IBM-SPSS for Windows version 22.0 software. The tests were performed with a significance level of 5%45,46,47. The Kolmogorov–Smirnov test was used to evaluate the distribution of the data of the studied sample. The values of the ECG and VCG parameters were expressed according to the normality tests, using mean and standard deviation or medians and quartiles. The parameters of the electrode positioning methods were compared using the paired Student’s t-test and Wilcoxon test. Regarding the qualitative data of the VCG loop, the results were described using absolute and relative frequencies, and the possible associations between the methods were analyzed using marginal homogeneity tests Kirkwood and Sterne45.