All the rabbits were anesthetized with urethane (1.2 g/kg) for induction. ECG monitor were then pasted to the animals. A left parasternal thoracotomy was performed, and the heart was exposed. The apex of one DAIG bipolar electrode (ST. JUDE company, America) was sutured to the epicardial surface of atrioventricular groove of left ventricular anterior wall, (Equivalent to the ventricular site of left anterior AP), and end of the electrode was kept outside through the cut, as ventricular pacing electrode. The other DIAG lead as atrial sensing electrode was placed at high right atrium via the internal jugular vein. The ends of the two leads were connected to the DF-4 electrophysiological stimulator (Oriental Electronic Company, Suzhou). Ventricular pacing was performed with voltage of twice the threshold of pulse amplitude, 2 ms of pulse width. We used a model ECG-6511 machine (Shanghai Medical Machine Company, China) to record surface ECG. Electrodes were placed on its arms and legs and right parasternal. The paper speed of ECG was 50 mm/s and the gain was 20 mm/mV. Programmed premature stimulation S₂ synchronized P wave positively swept ventricle (PS₂ interval: 0 → PJ interval, step length: 5 ms) to record surface ECG of chest lead, and the preexcitation model was simulated. Simultaneously, the ECGs were recorded. Programmed premature stimulation S₂ synchronized P wave positively swept ventricle (PS₂ interval: 0 → PJ interval, step length was 5 ms). The ECGs were recorded and preexcitation was simulated in the rabbit model. Cardioventricular pacing QRS complex by S₂ after T wave was recorded which represents the QRS induced by complete ventricular preexcitation.

(1) The ventricular activation time via normal pathway (sinus PR interval); (2) The initial, maximal and terminal QRS vector through normal pathway (sinus QRS); (3) The ventricular activation time via AP (PS₂ interval); (4) The initial, maximal and terminal QRS vector complete through AP (S₂ pacing QRS after T wave); (5) The initial, maximal and terminal QRS vector with AP conduction (R₂: Ventricular fusion wave formed by S₂ and sinus conduction).

(1) Initial QRS vector change: the direction or amplitude of initial 20 ms QRS vector which was different from normal pathway conduction; (2) maximal QRS vector change: the amplitude or direction of maximal QRS amplitude which was different from normal pathway conduction. (3) Terminal QRS vector change: the direction or amplitude of terminal 20 ms QRS vector which was different from normal pathway conduction.

(1) complete preexcitation: R₂ was absolutely equal to QRS complex through AP; (2) incomplete (typical) preexcitation: the initial vector of R₂ was equal to QRS complex through AP, but the maximal and terminal vector of QRS alters between normal and AP conduction (ventricular fusion wave); (3) incomplete latent preexcitation: R₂ was on behalf of ventricular fusion with terminal vector change; (4) complete latent preexcitation: R₂ was absolutely equal to QRS complex via normal pathway (S₂ encountered the effective refractory period of the ventricle).

Four kinds of different QRS waves were successfully observed on rabbit model of preexcitation, including complete preexcitation, incomplete (typical) preexcitation, incomplete latent preexcitation (terminal QRS vector change) and complete latent preexcitation (Fig. 1).

The relationships between the degree of ventricular preexcitation and the atrioventricular (AV) conduction time via A-V nodal pathway (PR interval), AV conduction time via AP (PS₂ interval), as well as the time difference of PR and PS₂ interval were shown in Table 1. AP conduction was faster than AV node conduction (PS₂ interval < PR interval), and R₂ was on behalf of overt preexcitation. When the difference was ≥47.00 ± 7.53 ms, R₂ was complete preexcitation, inversely, R₂ was incomplete preexcitation. AP conduction is not faster than AV node conduction (PS₂ interval ≥ PR interval), and R₂ is latent preexcitation. When the difference was ≤13.00 ± 3.50 ms, R₂ was incomplete latent preexcitation, inversely, R₂ was complete latent preexcitation.

The canine model of preexcitation syndrome has been reported in clinical ECG study [8]. In our rabbit model, sensing electrode was placed to high right atrium (sense sinus P wave) and stimulating electrode was placed to atrioventricular groove of left ventricular anterior wall (simulate preexcitation through a left anterior AP). Programmed premature S₂ which synchronized P wave (PS₂ interval was equal to the conduction time through AP) performed stimulation and built the animal model of preexcitation syndrome in which sinus P wave can simultaneously conduct through normal pathway and left anterior AP. Various degree of preexcitation (complete preexcitation, incomplete preexcitation, incomplete latent preexcitation and complete latent preexcitation) were obtained during changing PS₂ interval, confirming this study was a feasibility study. It can apply to the study of clinical ECG of preexcitation syndrome [9].

The preexcitation syndrome with AP capable of antegrade conduction could be divided into overt preexcitation and latent preexcitation. In latent preexcitation, the preexcitation is absent on the resting 12-lead ECG, which could be induced by transesophageal atrial pacing. Though there is no delta wave, the prevalence of arrhythmia and the risk of malignancy are same as the overt preexcitation [10‐13]. In this study we successfully imitated the ECG manifestation of 4 type preexcitation including complete ventricular preexcitation, incomplete ventricular preexcitation, incomplete latent preexcitation, complete latent preexciation. The finding of this study further confirms that the ECG manifestation of antegrade conduction of AP depends on the relative time that the atrial impulse is conducted down the AP into ventricle as opposed to the normal pathway. (1) When the AP conduction is faster than the AV node conduction, the overt preexcitation is observed on surface ECG. The degree of preexcitation depends on the how much faster the atrial impulse is conducted down the AP into ventricle as opposed to the normal pathway. If the relative time that the atrial impulse is conducted down the AP into ventricle as opposed to the normal pathway is longer than the time that impulse transmits from the ventricle that is directly connected to AP to the normal pathway (the time of this group is 47.00 ± 7.53 ms), the atrial impulse can’t conduct into ventricle via the normal pathway, resulting in the complete ventricular preexcitation. If the relative time that the atrial impulse is conducted down the AP into ventricle as opposed to the normal pathway is no longer than the time that impulse transmits from the ventricle that is directly connected to AP to the normal pathway, the impulse can conduct into ventricle by the normal pathway and AP forming the monophyletic ventricular fusion, resulting in incomplete ventricular preexcitation (Fig. 2). (2) When the AP conduction is slower than the AV node conduction, the delta wave is absent. If the relative time that the atrial impulse is conducted down the AP into ventricle as opposed to the normal pathway is no longer than the time that impulse transmits from the normal pathway the ventricle that is directly connected to AP (the time of this group is 13.00 ± 3.50 ms), the AP is still able to produce premature ventricular depolarization in the part of the ventricle that is directly connected to the AP forming the ventricular fusion, resulting in the change of terminal QRS vector. During this time the incomplete latent preexcitation could be seen in the ECG (Fig. 3). If the relative time that the atrial impulse is conducted down the AP into ventricle as opposed to the normal pathway is longer or equal to the time that impulse transmits from the normal pathway the ventricle that is directly connected to AP, the impulse can’t conduct into ventricle via AP (encounter the effective refractory period of the ventricle), showing complete latent preexcitation in the ECG.

In this study the overt preexcitation was divided into complete and incomplete ventricular preexcitation based on the ECG manifestation. The clinical significance is as follows: as to the incomplete ventricular preexcitation which is the classical manifestation of preexcitation syndrome, it is easy to identify; as to the complete ventricular preexcitation, it is easy to misdiagnose resulting from lack of recognition. (1) When AP closes to the pacemaker (atrial premature beats), the intra-atrial conduction time is significantly shortened, resulting in the superimposition of P′ wave and QRS complex. During this time the PR interval is indistinguishable, and the premature atrial contraction is often misdiagnosed as premature ventricular contraction [14]. Furthermore, the polarity of initial QRS vector is identical to the delta wave, the finding of premature P′ wave is helpful for the diagnosis. (2) When the AV node conduction is obviously slower than AP conduction, such as first or third degree atrioventricular block (AVB) in normal pathway with complete ventricular preexcitation, the AVB in normal pathway is masked which is usually misdiagnosed in clinic [15‐17]. During this time, the QRS complex is obviously widened and the PJ interval should be measured. The PJ interval prolongation provides that there is AVB in the normal pathway [15, 16, 18]. It is important for preoperative preparation and postoperative medical malpractice prevention to identify the presence of AVB according to ECG analysis before ablation.

In this study the latent preexcitation was divided into “incomplete latent preexcitation” and “complete latent preexcitation”. The clinical significance is as follows: the absence of delta wave is not means that the impulse can’t conduct down AP into ventricle. In patients with incomplete latent preexcitation syndrome, the delta wave is absent, but the impulse can conduct into ventricle via AP resulting in the change of terminal QRS vector [6]. The recognition of this new type preexcitation theoretically updates the recognition of delta wave. Meanwhile, it is helpful for the diagnosis of preexcitation syndrome mainly with a change of terminal QRS vector and the analysis of curative effect of bypass ablation [19‐21]. However, it is not easy to observe the change of terminal QRS vector which could be found in comparison to the ECG during atrioventricular reentrant tachycardia or after ablation.

The multiple leads were not synchronously detected, and it had a certain influence on the accurate analysis of the terminal QRS vector. It will be helpful to analyze the relationship between the terminal QRS vector and the AP location if the multiple leads can be simulated. Besides, in previous study [22‐24], researchers have observed the effects of ionic channel, gene, oxidative stress and inflammation on arrhythmias, and ventricular depolarization. Therefore, further studies are needed to assess whether these factors have influences on the electrocardiogram of antegradely conducting accessory pathway in future.