Abstract
It is difficult to differentiate the origins of focal atrial tachycardias (ATs) in adjacent structures by electrocardiography (ECG) alone. The aim of this study was to evaluate whether the clinical features of these ATs may help differentiate their origins.
One hundred and ninety-four patients (mean age, 43.5 ± 17.9 years; male, 53.6%) who underwent electrophysiological study for focal AT were included. We evaluated accuracy in differentiating the origin of AT by using ECG alone as well as with the addition of the clinical features. Electrocardiographs of ATs originating from the left superior pulmonary vein (LSPV, n = 24) vs. the left atrial appendage (LAA, n = 6), and from the right superior pulmonary vein (RSPV, n = 14) vs. the superior vena cava (SVC, n = 8) showed similar patterns. However, while no ATs from the LAA were found to be related to paroxysmal atrial fibrillation, 22 out of 24 ATs from the LSPV were associated with this condition. After localizing AT by using ECG, this clinical feature helped differentiate the ATs from the LSPV vs. the LAA with 93% accuracy. Moreover, while an on-and-off tachycardia (initiated and terminated more than 10 times per day) was observed in 4 of 8 ATs from the SVC, this pattern was observed in 13 of 14 ATs from the RSPV. After localizing the ATs by using ECG, on-and-off tachycardia helped differentiate the ATs from the RSPV vs. the SVC with 82% accuracy.
The clinical features and Holter monitoring can give additional information for differentiating the focal ATs originating from the adjacent structures.
We present a new electrocardiography algorithm combined with the clinical features for differentiating focal atrial tachycardia originating from the adjacent structures.
Introduction
Atrial tachycardia (AT) accounts for 7% of supraventricular tachycardias.1 Common origins of focal AT included the crista terminalis, tricuspid and mitral annuli, pulmonary veins (PV), coronary sinus (CS), interatrial septum, and right and left atrial appendages (RAA/LAA).2 Radiofrequency (RF) catheter ablation is a reliable treatment of focal AT, and the anatomical distribution of the foci producing these arrhythmias has been well characterized.2,3,4 This distribution can be recognized by P-wave morphology on electrocardiography (ECG).2 However, it is not easy to differentiate the origins of focal ATs from the adjacent structures by using ECG alone. In addition, P-wave is often superimposed on the T-wave and a complete isoelectric TP segment is not often appreciated in most clinical AT.
This study sought to evaluate whether an algorithm combined with the clinical features and the ECG characteristics could help differentiate the ATs originating from the left superior pulmonary vein (LSPV) vs. the LAA, and from the right superior pulmonary vein (RSPV) vs. the superior vena cava (SVC).
Methods
Patients
The study protocol was approved by the Institutional Review Board including the ethics committee of Severance Cardiovascular Hospital, Seoul, Korea and complied with the Declaration of Helsinki. The study population consisted of 194 consecutive patients (104 men and 90 women; mean age, 43.5 ± 17.9 years), who underwent an electrophysiological study and successful catheter ablation due to focal AT in the Yonsei University Health System between November 2002 and December 2011. The patients who had previously undergone catheter ablation due to any tachycardia were excluded. The mean duration from the onset of symptoms to an electrophysiological study was 1.8 ± 2.7 months. The mean duration from the ECG documentation of the AT to an electrophysiological study was 14.4 ± 11.2 days. Each patient provided written informed consent, and all antiarrhythmic drugs were discontinued for at least five half-lives prior to participation in the study.
Focal AT is defined as atrial activation starting rhythmically at a small area (a point source) from which it spreads out centrifugally and without endocardial activation over significant portions of the cycle length. In contrast, macroreentrant activation is conventionally defined as a circuit with a diameter of >2 cm and frequently occurs around a central obstacle. In focal AT, a discrete P-wave with an intervening isoelectric interval is observed. Conversely, in macroreentrant AT or atrial flutter, a continuous undulation without an isoelectric period on the ECG is demonstrated.5 Incessant AT was defined as AT which continued for 90% of the recording time. On-and-off AT was defined as AT that initiated and terminated more than 10 times per day on Holter monitoring. Tachycardia-induced cardiomyopathy was defined as heart failure with left ventricular ejection fraction ≤35% combined with incessant tachycardia, which was reversed after rate or rhythm control.
P-wave analysis for localization of focal atrial tachycardia
Twelve-lead ECGs were recorded by digital ECG instruments (General Electric Healthcare). The digital sampling rate was 500 samples per second. The low and high cut-off frequencies were 0.5 and 100 Hz, respectively. P-wave morphology was independently reviewed by the two observers.
P-wave morphology in each of the 12-ECG leads was classified as positive (+), negative (−), isoelectric, or biphasic. The amplitudes of positive P-waves were classified as either <0.1 mV (+) or ≥0.1 mV (2+). P-wave morphologies on the ECGs during AT were analysed for identification of the AT origin.2
Electrophysiological study
An electrophysiological study was performed in the fasting state under local anaesthesia with mild-to-moderate sedation. A quadripolar catheter was positioned across the tricuspid valve to record His-bundle activation. A second quadripolar catheter was positioned in the high right atrium (RA) or in the right ventricular apex as needed. A 7-Fr decapolar catheter was then introduced into the CS and advanced from the ostium to the distal portion of the CS (anterior portion of the mitral annulus adjacent to the LAA) to identify the earliest atrial activation within the CS. Left atrial (LA) catheterization was performed through either a patent foramen ovale or a transseptal puncture guided by intracardiac echocardiography. Intravenous heparin was administered to maintain an activated clotting time >250 s during left catheterization. Twelve-lead surface ECG and bipolar intracardiac electrogram recordings were displayed and stored on a computer-based digital amplifier system (Prucka Systems, General Electric Healthcare). Intracardiac electrograms were filtered between 30 and 500 Hz, and measurements were made offline by using digital calipers at a 200 or 400 mm/s sweep speed.
In the event that AT did not occur spontaneously, induction was attempted by using programmed electrical stimulation and/or isoproterenol infusion (1–2 μg/min). A diagnosis of AT was made by using standard electrophysiological criteria as follows: (i) tachycardia induction and maintenance independent of atrioventricular (AV) nodal conduction, and (ii) exclusion of accessory pathways, AV nodal reentry, sinoatrial reentrant tachycardia, and atrial flutter as mechanisms of arrhythmia.6 Detailed attention was given to the P-waves and atrial activation sequences of all the ATs to ensure that they were identical to the previously documented clinical AT.
Radiofrequency catheter ablation
To identify AT foci, electroanatomical mapping was performed in some patients as previously reported.7 In addition, standard activation mapping using the earliest atrial activation within the CS or P-wave onset was used as a reference in the remaining patients. After an origin of AT was identified, right or left atriogram was performed for exact anatomical information. Radiofrequency application with a target temperature of 60°C and maximum power output of 50 W using a 4 mm tip catheter (EP Technologies) or 30–40 W using a 3.5 mm irrigated-tip catheter (Celsius, Johnson & Johnson Inc.) was delivered at the presumed ablation site that exhibited the earliest bipolar activity and/or local unipolar QS pattern during AT. If the AT was unaffected after 10 s of RF catheter ablation, the RF application was terminated, and the catheter was repositioned. When the acceleration or the reduction of the AT was observed during the first 10 s of the energy application, RF delivery was continued for 30–60 s. The endpoint of RF catheter ablation was elimination and non-inducibility of the AT during isoproterenol infusion (2–4 μg/min) and burst atrial pacing to a cycle length as short as 300 ms.
Statistical analysis
Continuous variable data are reported as the mean ± standard deviation for normally distributed data, and as median (interquartile range) for abnormally distributed data. Categorical variable data are reported as count (percentage). We calculated the sensitivity, specificity, positive- and negative-predictive values, and accuracy of the combined ECG and clinical symptom approach for differentiating the ATs originating from the adjacent sites. Accuracy rate of the algorithm was calculated by the next formula; accuracy rate = (the number of cases of whom a specific AT origin was corrected by an algorithm)/(the number of all cases with the AT origin) × 100. Inter- and intra-observer repeatability for the P-wave morphology was calculated by using the kappa statistic. The Statistical Package for the Social Sciences version 20.0 (IBM Inc.) was used to perform all statistical evaluations. A P value of <0.05 was considered to be statistically significant.
Results
Distribution of the origins of atrial tachycardia
Figure 1 shows common AT foci for all 194 patients enroled in the study. Atrial tachycardia originating from the RA was more common than AT originating from the LA. Other common AT foci, in order of frequency, included the crista terminalis, os of the CS, LSPV, and RA septum.
The origins of focal AT and their frequencies. One red dot represents five patients. CS, coronary sinus; CT, crista terminalis; LA, left atrium; LAA, left atrial appendage; LIPV, left inferior pulmonary vein; LSPV, left superior pulmonary vein; MA, mitral annulus; RA, right atrium; RAA, right atrial appendage; RIPV, right inferior pulmonary vein; RSPV, right superior pulmonary vein; SVC, superior vena cava; and TA, tricuspid annulus.
Accuracy of the electrocardiography algorithms
The overall accuracy rates for localizing the origins of the focal ATs from our data by using ECG algorithms previously reported by Kistler et al.,2 Tang et al.,8 and Qian et al.9 were 78, 60, and 55%, respectively (Table 1). Discrimination between the ATs from the left PV vs. the LAA, and from the right PV vs. the SVC had low accuracy rates from 25 to 50%.
Accuracy rates according to AT origins on previously reported ECG algorithms
| Origin | Kistler et al.2 (%) | Tang et al.8 (%) | Qian et al.9 (%) |
|---|---|---|---|
| Crista terminalis | 86 | 87 | 50 |
| Os of coronary sinus | 86 | 71 | 71 |
| Right septum | 79 | 60 | – |
| Tricuspid annulus | 66 | 50 | 50 |
| LPV/LAA | 50 | 25 | 25 |
| RPV/SVC | 43 | 25 | 50 |
| Mitral annulus | 100 | 50 | 50 |
| Overall | 78 | 60 | 55 |
| Origin | Kistler et al.2 (%) | Tang et al.8 (%) | Qian et al.9 (%) |
|---|---|---|---|
| Crista terminalis | 86 | 87 | 50 |
| Os of coronary sinus | 86 | 71 | 71 |
| Right septum | 79 | 60 | – |
| Tricuspid annulus | 66 | 50 | 50 |
| LPV/LAA | 50 | 25 | 25 |
| RPV/SVC | 43 | 25 | 50 |
| Mitral annulus | 100 | 50 | 50 |
| Overall | 78 | 60 | 55 |
LAA, left atrial appendage; LPV, left pulmonary veins; RPV, right pulmonary veins; SVC, superior vena cava.
Accuracy rates according to AT origins on previously reported ECG algorithms
| Origin | Kistler et al.2 (%) | Tang et al.8 (%) | Qian et al.9 (%) |
|---|---|---|---|
| Crista terminalis | 86 | 87 | 50 |
| Os of coronary sinus | 86 | 71 | 71 |
| Right septum | 79 | 60 | – |
| Tricuspid annulus | 66 | 50 | 50 |
| LPV/LAA | 50 | 25 | 25 |
| RPV/SVC | 43 | 25 | 50 |
| Mitral annulus | 100 | 50 | 50 |
| Overall | 78 | 60 | 55 |
| Origin | Kistler et al.2 (%) | Tang et al.8 (%) | Qian et al.9 (%) |
|---|---|---|---|
| Crista terminalis | 86 | 87 | 50 |
| Os of coronary sinus | 86 | 71 | 71 |
| Right septum | 79 | 60 | – |
| Tricuspid annulus | 66 | 50 | 50 |
| LPV/LAA | 50 | 25 | 25 |
| RPV/SVC | 43 | 25 | 50 |
| Mitral annulus | 100 | 50 | 50 |
| Overall | 78 | 60 | 55 |
LAA, left atrial appendage; LPV, left pulmonary veins; RPV, right pulmonary veins; SVC, superior vena cava.
Figures 2 and 3 show the ECGs of the ATs originating from the LSPV and the LAA, respectively. However, P-wave morphology was positive in lead V1 and the inferior leads (II, III, and aVF), and negative in the lateral leads (I and aVL) in both cases.
Atrial tachycardia originating from the LSPV. (A) Twelve-lead ECG, (B) intracardiac electrogram, and (C) fluoroscopic images. The ablation (Abl) and lasso catheters were placed in the LSPV. The earliest signal was recorded in the LSPV. Red asterisks indicate the earliest potential and ablation site. The distal and the proximal parts duodecapolar (Halo, H) catheter was positioned from the distal CS to the RA. LAO, left anterior oblique; RAO, right anterior oblique; and SVC, superior vena cava.
Atrial tachycardia originating from the tip of the LAA. (A) Twelve-lead ECG, (B) intracardiac electrogram, (C) fluoroscopic images, and (D) three-dimensional electromagnetic activation map. The ablation catheter (Abl) was placed into the LAA. The earliest signal was recorded in the LAA. Red asterisks indicate the earliest potential and the ablation site. The distal and the proximal parts duodecapolar (Halo, H) catheter was positioned from the distal CS to the RA. LAO, left anterior oblique; RAO, right anterior oblique.
Figures 4 and 5 show the ECGs of the ATs originating from the RSPV and the SVC, respectively. P-waves were positive in lead V1 and the inferior leads, and negative in lead aVR in both cases. Biphasic P-wave in leads V1 and aVL was observed in AT originating from the SVC (Figure 4). In contrast, no biphasic P-wave in leads V1 or aVL was observed in patients with AT originating from the RSPV (Figure 5).
Atrial tachycardia originating from the RSPV. (A) Twelve-lead ECG, (B) intracardiac electrogram, and (C) fluoroscopic images. The ablation catheter (Abl) was placed into the RSPV. The earliest signal was recorded in the RSPV. Red asterisks indicate the earliest potential and the ablation site. The duodecapolar (Halo, H) catheter was positioned around the mitral annulus. CS, coronary sinus; LAO, left anterior oblique; RA, right atrium; RAO, right anterior oblique; and RV, right ventricle.
Atrial tachycardia originating from the SVC. (A) ECG, (B) intracardiac electrogram, and (C) fluoroscopic images. The ablation (Abl) and the lasso catheters (L) were placed into the SVC. The earliest signal was recorded in the SVC. Red asterisks indicate the earliest potential and the ablation site. CS, coronary sinus; LAO, left anterior oblique; RA, right atrium; and RAO, right anterior oblique.
Electrocardiography and the clinical features of atrial tachycardias originating from the adjacent structures
Table 2 shows the baseline characteristics of patients with AT originating from the LSPV (n = 24), LAA (n = 5), RSPV (n = 14), or SVC (n = 8). Among them, the numbers of patients with AT documented by 12-lead ECG were 22, 5, 9, and 3, respectively. P-wave morphology in the remaining patients could not be analysed because their ATs were documented by using a three-lead Holter monitor or a one-lead event recorder. Holter monitoring was performed in 28 of 51 patients. Among the patients with AT originating from the LSPV or the LAA, the percentage of patients with typical P-wave morphology (V1+, II/III/aVF+, I−, aVL− in AT) was 88.9%. Bifid P-wave in leads II or V1 could be observed in three cases (13.6%) with AT originating from the LSPV and four cases (80%) with AT originating from the LAA. Among the patients with AT originating from the RSPV or the SVC, the percentage of patients with typical P-wave morphology (V1+, II/III/aVF+, I+, aVR−, in AT; V1+ in sinus rhythm) was 83.3%. Biphasic P-wave in leads V1 or aVL was observed in 66.7% of patients with AT originating from the SVC. In contrast, no patients with AT originating from the RSPV had biphasic P-wave in leads V1 or aVL. Kappa statistic for inter- and intra-observer repeatability of P-wave morphology was 0.782.
Baseline characteristics of patients with AT of difficult-to-predict origin on the ECG algorithms
| AT origin | LSPV | LAA | P | RSPV | SVC | P |
|---|---|---|---|---|---|---|
| Number of patients | 24 | 5 | 14 | 8 | ||
| Number of patients with 12-lead ECGa | 22 | 5 | 9 | 3 | ||
| Age (years) | 57 (45, 65.25) | 25 (12, 47) | 0.027 | 44 (26.25, 50.5) | 34.5 (27.75, 42.5) | 0.837 |
| Male sex | 17 (70.8%) | 2 (40%) | 0.306 | 7 (50%) | 5 (62.5%) | 0.675 |
| P-wave | ||||||
| V1 (+) | 22 (100%) | 5 (100%) | 1.0 | 9 (100%) | 3 (100%) | 1.0 |
| Biphasic V1/aVL | 0 (0%) | 0 (0%) | 1.0 | 0 (0%) | 2 (66.7%) | 0.022 |
| II/III/aVF (+) | 22 (100%) | 5 (100%) | 1.0 | 7 (77.8%) | 3 (100%) | 0.378 |
| I/aVL (+) | 0 (0%) | 0 (0%) | 1.0 | 9 (100%) | 3 (100%) | 1.0 |
| I/aVL (−) | 19 (86.4%) | 5 (100%) | 0.553 | 0 (0%) | 0 (0%) | 1.0 |
| Clinical features | ||||||
| Incessant | 0 (0%) | 5 (100%) | <0.001 | 0 (0%) | 1 (12.5%) | 0.364 |
| Combined AF | 22 (91.7%) | 0 (0%) | <0.001 | 2 (14.3%) | 1 (12.5%) | 1.0 |
| On-and-off | 0 (0%) | 0 (0%) | 1.0 | 13 (92.9%) | 4 (50%) | 0.039 |
| TICMP | 1 (4.2%) | 2 (40%) | 0.068 | 0 (0%) | 0 (0%) | 1.0 |
| None | 2 (8.33%) | 0 (0%) | 0.511 | 1 (7.1%) | 3 (37.5%) | 0.083 |
| AT origin | LSPV | LAA | P | RSPV | SVC | P |
|---|---|---|---|---|---|---|
| Number of patients | 24 | 5 | 14 | 8 | ||
| Number of patients with 12-lead ECGa | 22 | 5 | 9 | 3 | ||
| Age (years) | 57 (45, 65.25) | 25 (12, 47) | 0.027 | 44 (26.25, 50.5) | 34.5 (27.75, 42.5) | 0.837 |
| Male sex | 17 (70.8%) | 2 (40%) | 0.306 | 7 (50%) | 5 (62.5%) | 0.675 |
| P-wave | ||||||
| V1 (+) | 22 (100%) | 5 (100%) | 1.0 | 9 (100%) | 3 (100%) | 1.0 |
| Biphasic V1/aVL | 0 (0%) | 0 (0%) | 1.0 | 0 (0%) | 2 (66.7%) | 0.022 |
| II/III/aVF (+) | 22 (100%) | 5 (100%) | 1.0 | 7 (77.8%) | 3 (100%) | 0.378 |
| I/aVL (+) | 0 (0%) | 0 (0%) | 1.0 | 9 (100%) | 3 (100%) | 1.0 |
| I/aVL (−) | 19 (86.4%) | 5 (100%) | 0.553 | 0 (0%) | 0 (0%) | 1.0 |
| Clinical features | ||||||
| Incessant | 0 (0%) | 5 (100%) | <0.001 | 0 (0%) | 1 (12.5%) | 0.364 |
| Combined AF | 22 (91.7%) | 0 (0%) | <0.001 | 2 (14.3%) | 1 (12.5%) | 1.0 |
| On-and-off | 0 (0%) | 0 (0%) | 1.0 | 13 (92.9%) | 4 (50%) | 0.039 |
| TICMP | 1 (4.2%) | 2 (40%) | 0.068 | 0 (0%) | 0 (0%) | 1.0 |
| None | 2 (8.33%) | 0 (0%) | 0.511 | 1 (7.1%) | 3 (37.5%) | 0.083 |
AF, atrial fibrillation; AT, atrial tachycardia; LAA, left atrial appendage; LSPV, left superior pulmonary vein; RSPV, right superior pulmonary vein; SVC, superior vena cava; TICMP, tachycardia-induced cardiomyopathy.
aThe number of patients with AT on 12-lead ECG.
Baseline characteristics of patients with AT of difficult-to-predict origin on the ECG algorithms
| AT origin | LSPV | LAA | P | RSPV | SVC | P |
|---|---|---|---|---|---|---|
| Number of patients | 24 | 5 | 14 | 8 | ||
| Number of patients with 12-lead ECGa | 22 | 5 | 9 | 3 | ||
| Age (years) | 57 (45, 65.25) | 25 (12, 47) | 0.027 | 44 (26.25, 50.5) | 34.5 (27.75, 42.5) | 0.837 |
| Male sex | 17 (70.8%) | 2 (40%) | 0.306 | 7 (50%) | 5 (62.5%) | 0.675 |
| P-wave | ||||||
| V1 (+) | 22 (100%) | 5 (100%) | 1.0 | 9 (100%) | 3 (100%) | 1.0 |
| Biphasic V1/aVL | 0 (0%) | 0 (0%) | 1.0 | 0 (0%) | 2 (66.7%) | 0.022 |
| II/III/aVF (+) | 22 (100%) | 5 (100%) | 1.0 | 7 (77.8%) | 3 (100%) | 0.378 |
| I/aVL (+) | 0 (0%) | 0 (0%) | 1.0 | 9 (100%) | 3 (100%) | 1.0 |
| I/aVL (−) | 19 (86.4%) | 5 (100%) | 0.553 | 0 (0%) | 0 (0%) | 1.0 |
| Clinical features | ||||||
| Incessant | 0 (0%) | 5 (100%) | <0.001 | 0 (0%) | 1 (12.5%) | 0.364 |
| Combined AF | 22 (91.7%) | 0 (0%) | <0.001 | 2 (14.3%) | 1 (12.5%) | 1.0 |
| On-and-off | 0 (0%) | 0 (0%) | 1.0 | 13 (92.9%) | 4 (50%) | 0.039 |
| TICMP | 1 (4.2%) | 2 (40%) | 0.068 | 0 (0%) | 0 (0%) | 1.0 |
| None | 2 (8.33%) | 0 (0%) | 0.511 | 1 (7.1%) | 3 (37.5%) | 0.083 |
| AT origin | LSPV | LAA | P | RSPV | SVC | P |
|---|---|---|---|---|---|---|
| Number of patients | 24 | 5 | 14 | 8 | ||
| Number of patients with 12-lead ECGa | 22 | 5 | 9 | 3 | ||
| Age (years) | 57 (45, 65.25) | 25 (12, 47) | 0.027 | 44 (26.25, 50.5) | 34.5 (27.75, 42.5) | 0.837 |
| Male sex | 17 (70.8%) | 2 (40%) | 0.306 | 7 (50%) | 5 (62.5%) | 0.675 |
| P-wave | ||||||
| V1 (+) | 22 (100%) | 5 (100%) | 1.0 | 9 (100%) | 3 (100%) | 1.0 |
| Biphasic V1/aVL | 0 (0%) | 0 (0%) | 1.0 | 0 (0%) | 2 (66.7%) | 0.022 |
| II/III/aVF (+) | 22 (100%) | 5 (100%) | 1.0 | 7 (77.8%) | 3 (100%) | 0.378 |
| I/aVL (+) | 0 (0%) | 0 (0%) | 1.0 | 9 (100%) | 3 (100%) | 1.0 |
| I/aVL (−) | 19 (86.4%) | 5 (100%) | 0.553 | 0 (0%) | 0 (0%) | 1.0 |
| Clinical features | ||||||
| Incessant | 0 (0%) | 5 (100%) | <0.001 | 0 (0%) | 1 (12.5%) | 0.364 |
| Combined AF | 22 (91.7%) | 0 (0%) | <0.001 | 2 (14.3%) | 1 (12.5%) | 1.0 |
| On-and-off | 0 (0%) | 0 (0%) | 1.0 | 13 (92.9%) | 4 (50%) | 0.039 |
| TICMP | 1 (4.2%) | 2 (40%) | 0.068 | 0 (0%) | 0 (0%) | 1.0 |
| None | 2 (8.33%) | 0 (0%) | 0.511 | 1 (7.1%) | 3 (37.5%) | 0.083 |
AF, atrial fibrillation; AT, atrial tachycardia; LAA, left atrial appendage; LSPV, left superior pulmonary vein; RSPV, right superior pulmonary vein; SVC, superior vena cava; TICMP, tachycardia-induced cardiomyopathy.
aThe number of patients with AT on 12-lead ECG.
While none of the patients with AT originating from the LAA exhibited paroxysmal atrial fibrillation (AF) at the time of diagnosis of AT, 22 (91.7%) of those with AT originating from the LSPV had paroxysmal AF at the time of diagnosis of AT. Atrial tachycardia originating from the LAA was incessant in five (100%) cases and was related to tachycardia-induced cardiomyopathy in two (40%) cases. On the other hand, none of the ATs originating from the LSPV were incessant. Only one patient (4.2%) with AT originating from the LSPV exhibited tachycardia-induced cardiomyopathy. Atrial tachycardias originating from the SVC and the RSPV exhibited on-and-off tachycardia in 4 (50.0%) and 13 (92.9%) cases, respectively.
Use of the electrocardiography algorithm in conjunction with the clinical features
We created an ECG algorithm by combining P-wave morphology with the clinical features for differentiating the origins of the focal ATs (Figure 6). Among the ATs originating from the LSPV or the LAA by the conventional ECG algorithm, incessant AT was considered as AT to have originated from the LAA. Atrial tachycardia combined with paroxysmal AF during Holter monitoring at the time of diagnosis of AT, on the other hand, was considered to have originated from the LSPV. Among the ATs originating from the RSPV and the SVC by the conventional ECG algorithm, AT with biphasic P-wave in leads V1 or aVL was considered to have originated from the SVC. In the next step, AT with on-and-off features was considered to have originated from the RSPV. Atrial tachycardia without on-and-off features was considered to have originated from the SVC.
The ECG algorithms combined with the clinical features for differentiating the origins of focal AT. AF, atrial fibrillation; AT, atrial tachycardia; LAA, left atrial appendage; LSPV, left superior pulmonary vein; RSPV, right superior pulmonary vein; SR, sinus rhythm; and SVC, superior vena cava.
The sensitivity, specificity, positive- and negative-predictive values, and the accuracy of the ECG algorithm combined with the clinical features for differentiating AT origin are shown in Figure 6. However, the conventional ECG algorithm could not differentiate between the ATs originating from the LSPV vs. the LAA or from the RSPV vs. the SVC.
Discussion
Main findings
The present study characterizes the anatomical distributions of focal ATs. We found that the clinical features provide valuable information in addition to that obtained via ECG regarding the origins of focal ATs. Clinical information such as the presence of incessant tachycardia or paroxysmal AF can help differentiate AT originating from the LSPV vs. the LAA. In addition, on-and-off tachycardia can help differentiate AT originating from the RSPV vs. the SVC.
Atrial tachycardias originating from the adjacent structures
According to the P-wave algorithm,2,8,9 the origin of AT can be predicted by analysing the P-wave morphologies on 12-lead ECG. The overall accuracy of this method ranges from 55 to 78%. It was difficult, however, to differentiate between ATs originating from the LSPV vs. the LAA, or from the SVC vs. the RSPV by using ECG due to the close anatomical proximity of these structures. In addition, it is not easy that AT is captured on a 12-lead ECG in clinical practice. Even if AT is captured on a 12-lead ECG, P-wave is often superimposed on the T-wave and a complete isoelectric TP segment is not often appreciated in most clinical AT. In these cases, the P-wave algorithm cannot be used. In 32% of AT, P-waves were difficult to delineate due to the P-wave on the T-wave from our data. After rate control, the P-waves could be differentiated from the T-waves.
Discrimination of atrial tachycardia originating from the left superior pulmonary vein vs. the left atrial appendage
For AT originating from the left PVs, the P-wave morphology was isoelectric or negative in lead I and bifid positive in leads II and/or V1.2 In the tachycardias arising from the LAA, P-wave morphology was similar to that originating from the left PVs. Kistler et al.2 suggested that the presence of a deeply negative P-wave in lead I was indicative of an origin in the LAA. On the other hand, Wang et al.10 reported that the P-waves were upright or biphasic (±) in lead V1 and isoelectric in leads V2–V6 during AT originating from the LAA. The upright deflection in lead V1 is probably due to the LA origin of the appendage, and the P-wave morphology observed in leads V2–V6 is because of the more anterior position of the LAA compared with the PVs. In this study, patients with ATs originating from the LSPV (n = 24) and LAA (n = 5) showed similar P-wave morphology. In the present study, bifid P-waves in leads II or V1 were not very useful for differentiating AT originating from the left PV and the LAA from AT with other origins.
The clinical features of the two types of ATs were different; however, the AT originating from the LSPV was frequently associated with paroxysmal AF. The accuracy rate of the algorithm was 93.1%, which was better than the previously published algorithms. Several studies have previously shown that the PVs are important sources of ectopic atrial beats or initiation of paroxysmal AF.11,12 Interestingly, AT from the LAA showed incessant tachycardia. These findings are consistent with the previous studies.10,13 Incessant tachycardia in cases of AT originating from the LAA may be due to enhanced automaticity.14 As a result, patients with AT originating from the LAA with incessant tachycardia may be at risk for tachycardia-induced cardiomyopathy.15 Left superior pulmonary vein isolation could be the procedure of choice rather than focal ablation in patients with AT originating from the LSPV combined with AF.
Atrial tachycardias originating from the right superior pulmonary vein and the superior vena cava
P-wave morphology during AT originating from the RSPV was positive in leads V1, I, II, III, and aVF.2 P-wave morphology during AT from the SVC, on the other hand, was highly positive in leads II, III, and aVF, and isoelectric in lead aVL. In addition, lead V1 showed a biphasic component.16 Kuo et al.17 reported that the combination of a biphasic or isoelectric P-wave polarity in lead V1 or a biphasic P-wave polarity in lead aVL had a sensitivity of 71%, specificity of 82%, positive-predictive value of 80%, and negative-predictive value of 74% for predicting an arrhythmogenic focus in the SVC. In addition, the notching and the angle of the P-wave could also be helpful in discriminating AT originating from the RSPV vs. the SVC.18,19 Comparison of the P-waves during tachycardia and during sinus rhythm can be the other useful method for distinguishing AT originating from the RSPV and the SVC.2
Among the eight patients with AT originating from the SVC, four patients presented with on-and-off AT unlike the algorithm. Therefore, specificity of the step with on-and-off feature in the algorithm was relatively low. However, we could improve the accuracy by adding ‘biphasic P-wave in leads V1 or aVL’ step. The accuracy rate of the algorithm was 81.8%, which was better than the previously published algorithms. The P-wave morphology in our study was similar to that reported by Zhao et al., although the clinical presentations described in the two studies were dissimilar. This difference may be due to inconsistencies in defining ‘frequent paroxysmal’ or ‘on-and-off tachycardia’.
We could not exactly explain why the AT originating with the LSPV were more associated with AF than the AT originating from the RSPV. It may be because the LSPV is more frequently AF-triggering focus than the RSPV, although both LSPV and RSPV are AF-triggering foci.11,20
Limitations
The number of the patients enroled in this study was relatively small. Further large-scale studies will be necessary to confirm the results reported here. In addition, the mechanisms of the focal ATs were not analysed. Elucidating these mechanisms in the future will allow us to understand why focal ATs from each site of origin present with specific clinical features. For finding out the mechanisms of tachycardia (reentry, triggered activity, or automaticity), mechanistic manoeuvres including induction and termination by programmed electric stimuli, and entrainment pacing could be important. Large-scale prospective studies are needed for validating this ECG algorithm combined with the clinical features.
Conclusions
Differentiation of the ATs based on the ECG criteria is difficult for distinguishing the origin from the adjacent structures, like the SVC from the RSPV or the LSPV from the LAA. The clinical features and Holter monitoring can give additional information for differentiating the focal ATs originating from such structures. A close review of the clinical features and Holter monitoring before electrophysiological study is important.
Funding
This work was supported in part by research grants of the Basic Science Research Program through the National Research Foundation of Korea, funded by the Ministry of Education, Science, and Technology (2012-0007604, 2012-045367), and a grant of the Korean Healthcare Technology R&D project, Ministry of Health and Welfare (A121668).
Acknowledgements
The authors thank Dong-Su Jang from the Department of Anatomy, Yonsei University College of Medicine for his medical illustration. The authors of this manuscript have verified that they comply with the Principles of Ethical Publishing in the International Journal of Cardiology.
Conflict of interest: none declared.
