• Open Access

Utility of 12-Lead Electrocardiogram for Differentiating Paroxysmal Supraventricular Tachycardias in Dogs


Corresponding author: Roberto A. Santilli, Dr Med Vet DECVIM-CA (Cardiology), Clinica Veterinaria Malpensa, Viale Marconi, 27, 21017 Samarate, Varese, Italy; e-mail: r.santilli@ecgontheweb.com.


Background: The 12-lead surface ECG is validated for differentiating supraventricular tachycardias (SVT) in humans. Despite the description of SVT in veterinary medicine, no studies have analyzed the electrocardiographic features of this type of arrhythmias in dogs.

Objective: To describe the specific electrocardiographic criteria used to differentiate the most common SVT in dogs.

Animals: Twenty-three dogs examined at Clinica Veterinaria Malpensa for SVT with the mechanism documented by electrophysiologic studies (EPS).

Methods: Twelve-lead electrocardiographic variables obtained from 14 dogs with orthodromic atrioventricular reciprocating tachycardia (OAVRT) and 9 dogs with focal atrial tachycardia (FAT) were compared.

Results: Dogs with FAT had faster heart rates (278 ± 62 versus 229 ± 42 bpm; P= .049) and less QRS alternans (33 versus 86%; P= .022). P waves appeared during tachycardia in 22 dogs, with a superior axis in 100% of OAVRT and 22% of FAT (P < .001). OAVRT was characterized by a shorter RP interval (85.0 ± 16.8 versus 157.1 ± 37.3 ms; P < .001) and smaller RP/PR ratio (0.60 ± 0.18 versus 1.45 ± 0.52; P < .001). Repolarization anomalies were present in 64% of OAVRT and no FAT (P < .001). Multivariate analysis identified QRS alternans and a positive P wave in aVR during tachycardia as independent predictors of arrhythmia type.

Conclusion and Clinical Importance: Electrocardiographic criteria used in people for differentiating SVT can also be applied in dogs.

The term supraventricular tachycardia (SVT) is commonly used to describe a group of arrhythmias involving the atrioventricular junction (atrioventricular nodal reciprocating tachycardia, AVNRT), the atrium (focal atrial tachycardia, FAT), or an extra nodal accessory pathway (orthodromic atrioventricular reciprocating tachycardia, OAVRT) in their circuit.1 The term supraventricular arrhythmias refers, instead, to sinus tachyarrhythmias, FAT, macro-reentrant atrial tachycardia (atrial flutter), AVNRT, OAVRT, focal, and nonparoxysmal junctional tachycardia excluding atrial fibrillation.1 OAVRT and incessant forms of atrial tachycardia are the only SVT reported in the dog.2–7 According to detailed endocardial mapping, electrophysiologic properties and anatomical disribution of accessory atrioventricular pathways responsible for OAVRT have been described in the dog.8 Both OAVRT and AVNRT are maintained by an anatomic reentrant circuit, whereas FAT can be caused by enhanced abnormal automatism, microreentrant circuits, and rarely by trigger activity.1 Despite different underlining mechanisms, these forms of atrial tachycardia are considered to all arise from 1 focus within the atrial chambers.1

Twelve-lead surface ECG is considered the 1st step in the diagnosis of SVT in humans, and several studies have validated its utility in differentiating narrow QRS complex tachycardia in patients that have undergone electrophysiologic examination for symptomatic tachyarrhythmias.9–19 Depending on the study, location of the P wave relative to the QRS complex during tachycardia, P wave axis in the frontal and horizontal plane, PR and RP interval duration, presence of QRS alternation, or repolarization anomalies were good discriminators of tachycardia type.9–19 Reported overall accuracy in differentiating SVT in humans by means of specific surface ECG criteria ranges between 75 and 97.8%.9–11,13–15,17–19 P wave location, pseudo r′ wave in lead V1, QRS alternans and ST depression > 2 mm in lead II or elevation ≥ 1–1.5 mm in lead aVR during tachycardia, and presence of ventricular pre-excitation during sinus rhythm are independent predictor of tachycardia mechanisms in human patients with symptomatic SVTs.9–11,13–15,18,19

Despite the existence of symptomatic SVT in dogs,20–22 to our knowledge no studies have analyzed the differences in electrocardiographic appearance of these tachyarrhythmias in this species. The aim of our study was therefore to validate the use of specific electrocardiographic markers as diagnostic tools for the differentiation of SVT documented in a group of dogs with known tachycardia mechanisms.

Material and Methods

Study Population

Twenty-three dogs referred to the electrophysiology laboratory of Clinica Veterinaria Malpensa with documented SVT between January 2005 and December 2006 were included in the study. The 23 dogs included 8 Labrador Retrievers, 5 Boxers, 1 Dogue de Bordeaux, 1 Beagle, 1 Newfoundland, 1 Bull Terrier, 1 Doberman Pinscher, 1 Great Dane, 1 Cavalier King Charles Spaniel, 1 Kurzhaar, 1 Neapolitan Mastiff, and 1 Dogo Argentino. Twenty-two were males of age 29.5 ± 31.6 (mean ± standard deviation [SD]) months and weight 31.1 ± 15.1 kg. Three dogs had not received any treatment, 9 dogs were being treated with quinidine (6 mg/kg PO q8h), 6 with metoprolol (0.25 mg/kg PO q12h), 2 with sotalol (0.35 mg/kg PO q8h), 1 with amiodarone (10 mg/kg PO q12h), 1 with verapamil (1 mg/kg PO q8h), and 1 with diltiazem (0.5 mg/kg PO q8h). All dogs underwent a physical examination, 12-lead surface ECG in right lateral recumbency using the described precordial position23 and interventional catheterization for endocardial mapping.

Electrophysiologic Study (EPS)

Before performing EPS, antiarrhythmic drugs were discontinued for at least 5 elimination half-lives.4,5 EPS was performed after preanesthetic medication with midazolana 0.2 mg/kg IV. General anesthesia was induced with propofolb 4 mg/kg IV bolus and maintained with a mixture of isofluoranec (1–2%) and oxygen (100%). Tachycardia mechanisms were determined according to the response to programmed atrial and ventricular stimulation and endocardial mapping of the arrhythmias circuit.5,8

Diagnostic criteria for OAVRT were the presence of eccentric ventricular-atrial activation, and in dogs with concentric ventricular-atrial activation it was differentiated from AVNRT when a single ventricular extrastimulus, introduced while the bundle of His was refractory, pre-excited the atrium with the same activation sequence or interrupted the tachycardia without reaching the atrium.5,8,24 Mapping of atrial or ventricular insertion of atrioventricular accessory pathways was performed as previously described.5,8 Entrainment of the tachycardia circuit was used to differentiate reciprocating tachycardia from FAT. After overdrive atrial pacing during tachycardia, the presence of variable (more than 10 ms from baseline) ventricular-atrial intervals after pacing was considered diagnostic for FAT.25 Reentrant atrial tachycardias were defined by entrainment with concealed fusion. Pacing 10–20 ms faster than the atrial tachycardia at a site with presystolic activity or middiastolic potential can entrain a reentrant tachycardia to the pacing rate without changing the morphology of P wave or the intracardiac electrogram sequence. Pacing in a dead-end site is excluded if the postpacing interval at the termination of concealed entrainment is identical to tachycardia cycle length.25,26 The atrial ectopic foci were localized as the site of earliest presystolic activity relative to the onset of the P wave during tachycardia where sharp and negative unipolar recording appeared.25 In case of difficulty determining the onset of P wave, delivery of a ventricular extrastimuli permitted advance of ventricular activation and repolarization and the distinction of ectopic P wave.24 The presence of disorganized atrial activity, characterized by irregular f waves with variable voltage, was considered diagnostic for atrial fibrillation.24 Electrical cardioversion with bipolar shock was used to restore sinus rhythm, when atrial fibrillation was induced during the study.


Twenty-three electrocardiographic recordings obtained during tachycardia and sinus rhythm were recorded for 2 minutes during EPS and analyzed by 1 author (RAS) with knowledge of the tachycardia type according to electrophysiologic testing results. Sinus rhythm strips were obtained after interruption of the tachycardia with programmed atrial or ventricular stimulation, chest thumps, or after ablation of the arrhythmic circuit or focus in case of incessant forms of tachycardia. Fourteen were OAVRT (63%) and 9 FAT (37%). Simultaneous 12-lead tracings were recorded at a paper speed of 50 mm/s with a gain setting of 10 mm/mV. The morphology of the P-QRS-T complexes during tachycardia was compared with the morphology during sinus rhythm and with the intracavitary signals.

Electrocardiographic Analysis

For each ECG, the following variables were analyzed: (1) QRS complex duration during tachycardia (in multiples of 20 ms); (2) heart rate during tachycardia; (3) QRS alternans during tachycardia, defined as a beat to beat variation in QRS amplitude of ≥1 mm in at least 1 lead (Fig 1A); (4) cycle length irregularity, defined as variation of tachycardia cycle length not on a beat to beat basis of ≥20 ms (Fig 1B) and cycle length alternans, defined as beat to beat oscillation of tachycardia cycle length of ≥20 ms; (5) presence of ventricular pre-excitation during sinus rhythm; (6) P wave configuration and polarity during tachycardia (P wave was considered visible if it appeared as a discrete deflection separate or within previous QRS-T complex); (7) RP interval duration (in multiples of 20 ms); (8) PR interval duration (in multiples of 20 ms); (9) RP/PR; (10) P wave axis in the frontal plane (defined as superior if negative in leads II and III, inferior if positive in leads II and III, intermediate if positive or biphasic in lead II and negative or biphasic in lead III, or indeterminate if the P wave axis could not be determined) (Fig 2); (11) P wave morphology and polarity in the horizontal plane; (12) atrioventricular relation during tachycardia; (13) repolarization abnormalities during tachycardia, including horizontal or down sloping ST segment depression of ≥2 mm in inferior leads that persisted 80 ms after J point; (14) a pseudo r′ wave deflection in lead V1 during tachycardia; and (15) pseudo S waves in leads II, III, and aVF during tachycardia.

Figure 1.

 (A) Rhythm strip in lead II in a dog with orthodromic atrioventricular reciprocating tachycardia. Notice the presence of QRS amplitude alternans on a beat to beat basis (paper speed 50 mm/s; 1 mV = 1 cm). (B) Rhythm strip in lead II in a dog with ectopic atrial tachycardia arising from the coronary sinus ostium. Notice the irregularity of cycle length with period of 2nd atrioventricular block (paper speed 50 mm/s; 0.5 mV = 1 cm).

Figure 2.

 Twelve-lead surface ECG (I, II, III, aVR, aVL, aVF, V1, V2, V3, V4, V5, V6) in 2 dogs during sinus rhythm and ectopic atrial tachycardia arising from the crista terminalis (A, B) and during sinus rhythm and orthodromic atrioventricular reciprocating tachycardia (OAVRT) (C, D). The comparison between sinus rhythm and tachycardia ECG allowed to locate P waves (arrows). P wave axis in the frontal plane is inferior for ectopic atrial tachycardia (B) and right superior for OAVRT (D) (paper speed 50 mm/s; 1 mV = 1 cm).

Statistical Analysis

Statistical analysis was performed with a freeware statistical software package (R 1.2.0).d Normally distributed data were ensured using the Shapiro-Wilk Normality Test. Metric data are presented as mean ± SD, and nominal data are expressed as a percentage. Normally distributed data were tested for statistical significance using unpaired Students t-test. Nonnormally distributed data were tested using Wilcoxon's sum rank test.

Univariate analysis of nominal data was performed with contingency table analysis by Fishers exact test or χ2 statistic if appropriate.

Variables that were statistically significant (P≤ .05 and odds ratio confidence intervals of 95% excluding 1) were considered relevant for a multivariate analysis by a stepwise logistic regression technique, having the tachycardia type as the dependent variables and the ECG criteria as the independent variables. The most parsimonial final model was selected, via backward elimination, with a Wald P-value of .05 as removal threshold, given an acceptable log-likelihood ratio test value. Model fit was evaluated by Pearson's and Hosmer-Lemeshows goodness-of-fit test.


QRS complex duration was similar during both narrow QRS complex tachycardia, 61.1 ± 3.3 ms in case of FAT and 60.7 ± 10.7 ms in dogs with OAVRT (P= .9). Dogs with focal AT had faster heart rates than those with OAVRT (278 ± 62 and 229 ± 42 bpm, respectively; P= .049) and less QRS alternans (33 and 86%, respectively; P= .022). QRS alternans was most likely to occur in limb leads II, III, and aVF and precordial leads V3 to V6. Despite the commonly seen cycle length irregularity during FAT (67%) and cycle length alternans during OAVRT (14%), there were no statistically significant differences. One dog with OAVRT and cycle length alternans had multiple accessory atrioventricular pathways. Ventricular pre-excitation during sinus rhythm was present in 3 dogs with OAVRT (21%) and in 1 dog (11%) with FAT (P= 1). P waves were detected in all FAT and in 13 out of 14 dogs with OAVRT (93%) (P= .12). RP interval was shorter during OAVRT than FAT (85.0 ± 16.8 and 157.1 ± 37.3 ms, respectively; P < .001), whereas PR was similar (120.7 ± 47.6 during FAT; 150.0 ± 38.3 during OAVRT; P= .17). RP/PR ratio was smaller in dogs with OAVRT than with FAT (0.60 ± 0.18 and 1.45 ± 0.52, respectively; P < .001). P wave axis in the frontal plane during tachycardia was superior in 100% of OAVRT (Fig 2D) and 22% of FAT (P < .001) and inferior in the remaining dogs (Fig 2B). All OAVRT and 7 FAT (78%) had 1 : 1 atrioventricular relation during tachycardia paroxysms (P= .15). Sixty-four percent of OAVRT and no FAT had repolarization anomalies (P < .001) (Table 1). Pseudo r′ wave in lead V1 and pseudo Q or S wave in inferior leads were not detected. Table 2 shows the polarity of P wave during tachycardia in standard and precordial leads.

Table 1.   Electrocardiographic findings in simultaneous 12 lead surface ECG during orthodromic atrioventricular tachycardia and focal atrial tachycardia in 23 dogs.
(n = 9)
(n = 14)
P Value
  1. OAVRT, orthodromic atrioventricular reciprocating tachycardia; FAT, focal atrial tachycardia; NS, not significant; CL, cycle length.

QRS complex duration (ms)61.1 ± 3.360.7 ± 10.7NS
Heart rate (bpm)278 ± 62229 ± 42.049
Irregular CL or CL alternans6/9 (67%)2/14 (14%)NS
QRS alternans3/9 (33%)12/14 (86%).022
Pre-excitation during sinus rhythm1/9 (11%)3/14 (21%)NS
Visible P wave9/9 (100%)13/14 (93%)NS
RP interval (ms)157.1 ± 37.385.0 ± 16.8< .001
PR interval (ms)120.7 ± 47.6150.0 ± 38.3NS
RP/PR1.45 ± 0.520.60 ± 0.18< .001
Superior P wave axis frontal plane2/9 (22%)13/13 (100%)< .001
1 : 1 atrioventricular relation7/9 (78%)13/13 (100%)NS
ST shift in inferior leads0/9 (0%)9/14 (64%)< .001
Table 2.   P wave polarity in simultaneous 12 lead surface ECG during orthodromic atrioventricular tachycardia and focal atrial tachycardia in 23 dogs.
LeadFAT (n = 9)OAVRT (n = 13)P Value
  1. OAVRT, orthodromic atrioventricular reciprocating tachycardia; FAT, focal atrial tachycardia; NS, not significant.

IPositive 7/9 (78%)Positive 8/13 (62%)NS
Negative 1/9 (11%)Negative 3/13 (23%) 
Biphasic 1/9(11%)Biphasic 2/13 (15%) 
IIPositive 7/9 (78%)Positive 0/13 (0%)< .001
Negative 2/9 (22%)Negative 13/13 (100%) 
IIIPositive 7/9 (78%)Positive 0/13(0%)< .001
Negative 2/9 (22%)Negative 13/13 (100%) 
aVRPositive 2/9 (22%)Positive 11/13 (85%)< .001
Negative 7/9 (78%)Negative 2/13 (15%) 
aVLPositive 5/9 (56%)Positive 13/13 (100%).02
Negative 3/9 (33%)Negative 0/13 (0%) 
Biphasic 1/9 (11%)Biphasic 0/13 (0%) 
aVFPositive 7/9 (78%)Positive 0/13 (0%)< .001
Negative 2/9 (22%)Negative 13/13 (100%) 
V1Positive 8/9 (89%)Positive 3/13 (23%)NS
Negative 1/9 (11%)Negative 10/13 (77%) 
V2Positive 7/9 (78%)Positive 4/13 (31%)NS
Negative 2/9 (22%)Negative 9/13 (69%) 
V3Positive 8/9 (89%)Positive 5/13 (38%).03
Negative 1/9 (11%)Negative 8/13 (62%) 
V4Positive 8/9 (89%)Positive 4/13 (31%).02
Negative 1/9 (11%)Negative 7/13 (54%) 
Biphasic 0/9 (0%)Biphasic 2/13 (15%) 
V5Positive 8/9 (89%)Positive 3/13 (23%)< .001
Negative 1/9 (11%)Negative 9/13 (69%) 
Biphasic 0/9 (0%)Biphasic 1/13(8%) 
V6Positive 8/9 (89%)Positive 3/13 (23%)< .001
Negative 1/9 (11%)Negative 9/13 (69%) 
Biphasic 0/9 (0%)Biphasic 1/13 (8%) 

The independent predictors of tachycardia type by multivariate analysis were the presence of QRS amplitude alternans (1/OR = 27) and a positive P wave in aVR (1/OR = 22) during OAVRT.


This study describes the utility of 12-lead surface ECG in distinguishing common SVT in dogs with narrow QRS complex tachycardias of known origin. Several electrocardiographic findings provided valid discriminators of tachycardia type, and different electrocardiographic reference ranges for SVTs in dog have been set. Despite this fact, the small size of the study might have affected the criteria asserted, and this should be considered when considering the results.

The duration of the QRS complex during both types of tachycardia was <70 ms, the upper limit for QRS complex duration during sinus rhythm in most breeds of dogs.27–29 Because antegrade conduction in both OAVRT and FAT is along the atrioventricular-His-Purkinje system, the QRS complex duration should remain within normal limits in the absence of functional or anatomical conduction abnormalities.24 In humans, a tachyarrhythmia is classified as a narrow QRS complex tachycardia when the QRS duration is < 11011,12,14,18 or 120 ms.9,10,13,15,16,19 According to our study, in dogs a tachycardia could be considered narrow QRS complex tachycardia if the QRS duration is < 70 ms.

In our report the ventricular rate was faster during FAT. This result is in contrast to what has been reported in humans where tachycardia cycle length was shorter in OAVRT10,18 or similar in all types of tachycardia analyzed.9,11,12,14,19 The small size of the present study, the decreased sympathetic tone present during anesthesia, and particularly the anesthetic agents used might have influenced the heart rate, altering the conduction properties of the His-Purkinje system and of the accessory pathways. However, because the PR interval, representing the atrioventricular conduction time, was not different between arrhythmias, the difference in heart rate found in our group of animals can only be attributable to a higher discharge rate of the atrial ectopic focus compared with the length of macroreentrant circuit of OAVRT. Similar to the humans, heart rate was not predictive of tachycardia type in our dogs.9,11,12,14,19

In our study, QRS alternans was most commonly seen during orthodromic AVRT and was an independent predictor of this arrhythmias. QRS alternans was most likely to occur in the same limb and precordial leads reported in the human literature.9,10 In humans, some authors found QRS alternans to be predictive of AVRT, particularly in patients without ST segment depression,9,11,14,16,30 suggesting that patients with tachycardia because of an accessory atrioventricular pathway may have anatomically or functionally different conduction systems than patients with other SVTs.30 This hypothesis was not confirmed by other authors who demonstrated that QRS alternans is a rate-related phenomenon that depends on an abrupt increase to a critical heart rate.10,19,31–33 According to these studies, the relative refractory period of the His-Purkinje system oscillates on a beat to beat basis. When a critical rate is reached suddenly, QRS alternans is caused by an alternation in the action potential duration because of oscillation in the diastolic interval.31–33 Other factors that may influence the R wave amplitude are changes in heart volume, ischemia, contractility, and electrical axis. Whether one or more of these factors contribute to QRS alternans during SVT is unknown.34 In contrast to the preceding data, our group of dogs with FAT had higher heart rates but only 33% had QRS alternans, which was not dependent on heart rate.

As reported in humans, cycle length alternans was not a discriminator of tachycardia mechanism in our study.10,14,15,19 Cycle length irregularity was more common during FAT, and it was probably caused by an unpredictable variation in the discharge rate of the ectopic focus, atrioventricular blocks, or dual atrioventricular nodal physiology.24,35 Orthodromic AVRT may have cycle length alternans in cases of multiple accessory pathways with alternating conduction, as happened in 1 dog of our study, or dual atrioventricular nodal physiology.24,35 A further explanation of cycle length alternans in this type of tachyarrhythmia is given by an experimental study conducted in a canine model of AVRT that showed the occurrence, amplitude, and duration of cycle length alternans are predictable based on atrioventricular node recovery properties and depend on retrograde conduction properties of the reentrant circuit.36

Ventricular pre-excitation was present only in one third of dogs with OAVRT.8 The majority of dogs with circus movement tachycardia presented accessory pathways with unidirectional retrograde conduction. These data are in contrast to what is reported in humans where accessory pathways have bidirectional conduction with overt signs of ventricular pre-excitation during sinus rhythm in 53–94% of cases.37–45 One dog with FAT in our study had intermittent ventricular pre-excitation along an accessory pathway with unidirectional antegrade conduction and inducible atrial tachycardia with rapid atrial pacing during EPS. This type of ventricular pre-excitation has been described in dog, and it can be detected on surface ECG only by prolonged monitoring.4 In humans, accessory pathways present unidirectional antegrade conduction in 0.8–6% of cases and no association with FAT has been reported.11,14,37–45

Contrary to human patients where multivariate analysis found ventricular pre-excitation as an independent predictor of tachycardia mechanism when comparing AVRT, AVNRT, and FAT,11,14 ventricular pre-excitation did not discriminate the tachycardia type in our study, although its presence lends support to OAVRT.

P wave visibility depends on the duration of the QRS complex and the ventriculo-atrial conduction time during OAVRT and the discharge rate of the atrial focus or the presence of atrioventricular block during FAT. The comparison of ECG during sinus rhythm and tachycardia, as reported,9 permitted us to locate P waves in all dogs with FAT and in 93% of OAVRT (Fig 2). The shortest ventriculo-atrial conduction time reported in children during AVRT is 60 ms46 whereas in dogs it is 40–60 ms.47 In our study, we did not measure this interval, but we attempted to use the RP interval obtained from surface ECG as an index of ventriculo-atrial conduction time.15 In humans, the P wave location permits discrimination among OAVRT, AVNRT, and FAT.9–11,13–17,19 Similar to our results, P waves were visible in 68–100% of OAVRT, 75–80% of FAT, and 32–70% of cases of AVNRT.9–11,13–17,19 In the last case, because of the shorter ventriculo-atrial conduction time, P waves were often buried or distorted previous QRS complexes and appeared as a pseudo r′ wave in lead V1 in 42–63% of patients and as a pseudo S wave in inferior leads in 7–20% of cases.9–11,13–17,19 This appearance of incomplete right bundle block in lead V1 during tachycardia was also reported in 7.5–8% of OAVRT and 8–10% of FAT, whereas no pseudo S waves in inferior leads were reported.11,14 One study found P wave location during tachycardia to be related to age and not to gender, with a higher proportion of AVNRT with visible P waves in elderly patients.17 The visibility of P waves in this population was caused by a delay in nodal retrograde conduction time in 77% of cases.17 In our study, neither OAVRT or FAT, despite the shorter ventriculo-atrial conduction interval reported in dogs, presented pseudo r′ wave in lead V1 or pseudo S wave in II, III, and aVF.

The RP interval on the surface ECG is often used to measure ventriculo-atrial conduction time during AVRT and AVNRT, and despite the ventriculo-atrial dissociation present during FAT, to differentiate SVT in humans.9–11,13–16,19 In our study, the RP interval was significantly shorter than PR, with a smaller RP/PR ratio during OAVRT than during FAT. Similar to our results, studies in humans have found short RP in 70–95% of AVRT and 10–33% of FAT.9–12,14–16,19 An RP interval ≥100 ms correctly discriminated patients with AVRT from AVNRT in 84–93% of cases, whereas a RP/PR ≥ 1 differentiated FAT from AVRT and AVNRT in 38% of patients.11,15,16,19 A difference of RP intervals in leads V1 and III ≥20 ms correctively differentiated posterior-type AVNRT from AVRT by means of a concealed postero-septal pathway with a sensitivity of 71%, a specificity of 87%, and positive predictive value of 75%.13

A right superior P wave axis in the frontal plane independently discriminated OAVRT from FAT in our group of dogs. The precordial lead system used in this study was particularly helpful for the identification of P waves during tachycardia and QRS amplitude alternans, but the analysis of the P waves axis in the horizontal plane did not discriminate the tachycardia mechanism. Breed variation of the heart position in the thorax compared with human beings and different locations of accessory pathways and atrial ectopic foci may explain the variable appearance of the P waves in precordial leads we found. In human patients, negative P wave polarity in inferior leads (II, III, and aVF) with positive aVR, aVL, were found in 26% of anterior type of AVNRT, 100% of cases of posterior type of AVNRT, 64% of AVRT, and 30% of FAT.9,13 P wave polarity in the horizontal plane showed a right to left appearance (negative P wave in lead V1 and positive in V6) in 84% of AVRT and 55% of FAT, and a left to right appearance (positive P wave in lead V1 and negative in V6) in 20% of anterior type of AVNRT.9 Unclear polarity was found in 80% of anterior type AVNRT.9 The P wave axis could not differentiate AVRT from AVNRT in a study because it was indeterminable in 73 and 44% of cases, respectively.10 Retrograde P wave polarity during tachycardia in leads V1, II, III, aVF, and I during AVRT permitted correct localization of concealed accessory pathways along the atrioventricular groove in humans, with an accuracy ranging from 75% for right midseptal bypass tracts to 93.8% for left posterior ones.13 According to the results of this study, retrograde P waves appear negative in the inferior leads in cases of infero-posterior free wall localization, isoelectric or biphasic in cases of lateral free wall localization, and positive in cases of superior-anterior wall localization. Lead I presents negative retrograde P waves in cases of left free wall, whereas positive or isoelectric P wave are characteristic of right free wall accessory pathways.13 Retrograde P wave during circus movement tachycardia mediated by postero-septal accessory pathways causes negative polarity in inferior leads, positive polarity in lead aVR and aVL, and isoelectric of biphasic in lead I.13 To further emphasize these results, a study of Waldo et al48 showed that negative P waves in inferior leads appeared when pacing the region of the coronary ostium, and they became positive or biphasic when pacing the atria from sites of the atrioventricular junction located more anteriorly. Negative P wave polarity in aVR during tachycardia identified crista terminalis ATs in humans with a sensitivity of 100% and a specificity of 93% and differentiated them from tricuspid annulus AT and septal AT, whereas positive P wave polarity in the inferior leads differentiated superolateral AT from inferolateral AT with a sensitivity of 86% and a specificity of 100%.49 In our study, most of the accessory pathways that mediated OAVRT were in the posterior and postero-septal region of the tricuspid valve annulus. This explains the high percentage of retrograde P wave superior axis with positive polarity in aVR found during OAVRT. Approximately two thirds of dogs with FAT had inferior P wave axis with negative P wave polarity in aVR during tachycardia, and the most common site of origin was the superolateral right atrium area. Positive P wave in lead aVR causing ST segment elevation has been found to be the only factor to differentiate narrow QRS complex tachycardia in a human study, with a sensitivity, specificity, and accuracy to discriminate AVRT from FAT and AVNRT of 71, 70, and 70 respectively.18 The value of lead aVR in clinical cardiology has been pointed out particularly during acute coronary syndromes such as during acute left main coronary artery obstruction, acute pulmonary embolism, and SVT with a positive retrograde P wave appearance in case of AVNRT and OAVRT using para-septally located accessory pathways.50

In our study, all dogs with OAVRT had 1 : 1 atrioventricular relation during tachycardia paroxysms, whereas 1 out of 3 dogs with FAT had variable degree of atrioventricular block. These data are in accordance with the human literature where 30% of AT presented 2nd degree atrioventricular block (Fig 1B).9 AVNRT typically has a 1 : 1 AV and VA relation, but rarely there can be retrograde block from the retrograde AV nodal pathway to the atrial myocardium.24

In humans, it has been shown that ST shift in inferior leads during narrow QRS complex tachycardia discriminated AVNRT from AVRT with a positive predictive value from 43 to 100% and from 54 to 77%, respectively.14–16,19 Inferior leads ST shift has been described in 33% of cases of FAT and ST elevation in aVR in 16%.18 In our study, we classified repolarization abnormalities only as inferior ST shift because ST elevation in aVR was considered as a retrograde P wave. As reported in the human literature we found inferior ST shift in 64.29% of dogs with OAVRT and in no dogs with FAT. Repolarization anomalies are often found during tachycardia and, even when marked, are usually not associated with myocardial ischemia because myocardial lactate extraction remained unchanged from baseline in 88% of patients.51 Possible mechanisms postulated may be a true ST shift produced by current flow only during systole caused by differences in action potentials in different regions of the heart,52,53 retrograde atrial activation with ST distortion in patients with AVRT,12,14–16,18,19 and catecholaminergic alteration of phase 2 of the action potential, which may result in ST shift.51

Limitations of the Study

This study included only dogs referred to our institution for EPS, and therefore this series might not be representative of all narrow QRS complex tachycardias in dogs. Because of the difficulty diagnosing SVT in veterinary medicine, the number of ECG included in the study was small and this might have influenced the results. In particular the small number of FAT did not allow us to analyze the electrocardiographic criteria according to the position of accessory pathway or atrial ectopic focus. All FAT analyzed in our study arose from the right atrium, and this is likely to positively influence the results because some left-sided FAT, although not yet reported in dogs, would produce right superior P wave axis shift similarly to posterior and postero-septal accessory pathways mediated OAVRT.13–18 In our sample only 1 anteroseptal accessory pathway maintained a circus movement tachycardia with left inferior retrograde P wave axis. A bigger sample, including more atrioventricular bypass tract locations, might mislead the diagnosis of OAVRT. A major limitation of the study was the fact that the electrocardiographic tracings during tachycardia were obtained in anesthetized dogs, and this might have accounted for a reduction of heart rate and prolongation of atrioventricular and ventriculo-atrial conduction times. Also the necessity to compare the electrocardiographic appearance of P-QRS-T complex during tachycardia with sinus rhythm limits the use of the criteria found. This was particularly important to establish P wave location and axis, RP interval, and ST shift, criteria that permitted us to discriminate tachycardia mechanisms. In case sinus rhythm strips are not available, P waves buried within the QRS complex, overt ventricular pre-excitation, and repolarization anomalies may not be seen. The same limitation has been found in different human studies.9,11,13,19 In our study, when sinus rhythm strips were not available, we attempted a chest thump during sustained tachycardia in order to have at least 1 or 2 sinus beats. This procedure allowed us to recognize P waves in all but 1 case, and therefore it can be used in a clinical setting. The precordial lead system used in this study was useful to analyze QRS alternans and to locate P waves but did not permit to differentiate OAVRT from FAT studying the P wave axis in the horizontal plane. Therefore different precordial lead systems should be tested during tachycardia to try to differentiate right from left AVRT and FAT subtypes.12,27–29,49,54


To our knowledge, this is the 1st study that analyzed the electrocardiographic appearance of narrow QRS complex tachycardia in veterinary medicine. Based on our results, some electrocardiographic criteria used in people for differentiating SVT can be applied also in dogs. Future prospective studies are needed to validate these findings, and stepwise algorithms should be formulated and tested to guide the clinicians in the diagnosis and the choice of treatment of complex supraventricular arrhythmias.


aMidazolan, PHT Pharma, Milan, Italy

bRapinovet, Schering-Plough, Milan, Italy

cIsoflurane, Merial Italia Spa, Assago, Milan, Italy

dR development core team (2005). R: Language and environment for statistical computing. R Foundation for statistical computing, Vienna, Austria, ISBN 3-900051-07-0


We thank Davide Sansottera and Giuliano Villa (Precise Med) for equipment and technical assistance during the procedures. Dr Barret Bulmer for reviewing the manuscript.