INTRODUCTION

Circumferential pulmonary vein (PV) isolation has proven to be an effective treatment for atrial fibrillation.1 Radiofrequency lesions are created in the transition zone between the ostia of the PV (PVO) and the left atrium (LA), avoiding stenosis of the PV,2 in addition to eliminating rotors, drivers, and lymph plexuses occupying this region.3 It is known that impedances (RB) of the PV (RPV), their ostia, and the LA (RLA) are different,4-7 and consequently the transition zone could be identified using 3-dimensional impedance mapping to guide ablation safely.

METHODS

The study included 25 patients (15 in the baseline study and 10 in the prospective phase) who underwent ablation for paroxysmal atrial fibrillation (Table 1). Transesophageal echocardiogram and computerized tomography of the chest were performed. The study was done with the patients under sedation with remifentanil and anticoagulation with sodium heparin. By means of a transseptal puncture a circular catheter was inserted into the LA to sequentially map each PV (Lasso, Biosense-Webster) with a 3.5 mm irrigated tip ablation catheter (Navi-Star, ThermocoolTM, Biosense-Webster).

Construction of the Impedance Map

The impedance circuit was composed of the radiofrequency generator (Stockert, Biosense-Webster), the ablation catheter and the reference patch (Ref-Star, Biosense-Webster) placed on all the patients between the second and sixth thoracic vertebrae. The RB was measured after the generator delivered a 50 kHz low amplitude (2 µA) electric current. Contact between the catheter and the tissue was considered to be optimal if the movement of the catheter and the electric signal were stable. For each point examined, the RB was measured at least twice before being recorded. To be considered valid it could not vary by more than 2 W in the same position.

Three-dimensional impedance mapping of the LA and the PV was carried out with an electroanatomical mapping system (CARTO XP, Biosense-Webster) with color-coding (red, the lowest value; violet, the highest) of the RB. Each PV and the LA were reconstructed on independent maps. With the ablation catheter, points were acquired from the interior of each PV (1-3 cm) up to the PVO, where points were taken along the entire circumference. The PVO was defined by the simultaneous appearance of an electric signal corresponding to the PV potential together with the LA potential in the distal electrode, and the absence of the PV potential in the proximal electrode, coinciding with the penetration of the catheter into the cardiac silhouette. It was considered to be in the LA when the appearance of the electric signal of the LA in the distal electrode without PV potential coincided with the characteristic movement of the catheter on exiting the vein. In the LA, points were taken on the posterior wall, roof, septum, mitral annulus, and appendage, taking special care and acquiring the greatest number of points possible in the vicinity of the PVO. The map obtained was incorporated into the computerized tomography image (CARTO MERGE Biosense-Webster) (Figure 1).

Figure 1. A: impedance map integrated into the computerized tomography (CARTO MERGE). B-C: diagram of 36 points acquired around the pulmonary veins (PV). LIPV indicates left inferior PV; LSPV, left superior PV; RIPV, right inferior PV; RSPV: right superior PV

Prospective Phase. Ablation

In the last 10 patients, an independent operator, blinded to the location of the catheter in the CARTO, classified 36 points per patient (Figure 1) as LA or PVO reading only the RB value. An RLA + 4% W (see below) was considered to be the impedance limit of the LA. Another operator who did not know the RB value classified the same point as LA or PVO with the CARTO. In all patients a circumferential ablation of the PV was performed by making an encirclement or circlet of ipsolateral veins (to an atrial electrogram voltage reduction of up to 90% or <0.05 mV) according to the impedance lines of the map (Figure 2), with radiofrequency applications of 35 W and a maximum temperature of 45°C. In the event of "distortion" of the impedance line, the anatomic image was used to select theapplication. If isolation of the PV was not achieved with the encirclement, confirmed by the disappearance of all the potentials in each PV with the Lasso catheter, energy was applied in the interior of the circlet until the aim of the ablation was achieved, which was isolation of all the PV. All patients were discharged with antiarrhythmia medication for 3 months and a control cardiac computerized tomography was performed.

Figure 2. A-D: projections of a 3-dimensional impedance map with a range between 133 ¿, lower limit in the left atrium (LA) (red), and 166 ¿ (violet) that is the superior limit obtained 1-3 cm within the pulmonary veins (PV). Impedance gradient from deep PV to LA stands out, as does the LA-PV transition (red-green). E-H: ablation (red dots) following the impedance lines in the PV-LA junction.

The statistical package SPSS14 (Chicago, IL) for Windows was used. The Student t test was used to identify significant differences between impedance means. The receiver operating characteristic (ROC) curve was drawn to graphically show the sensitivity and specificity of a particular impedance value that differentiated LA from PVO.

RESULTS

In the first 15 patients an RB gradient was observed from the interior of the PV (1-3 cm) to the LA: RPV, 165.4 (7.5) W; RPV ostium, 141.6 (7.3) W; RLA, 131.09 (8.3) W (Table 2), which was maintained individually in each patient (Figure 3A). The RLA showed no differences in the various zones analyzed (septum, mitral annulus, posterior wall, and roof), but was higher at the base of the left atrial appendage and at its vertex (Table 2). A value of 136 W (RLA + 4% W) defined the limits of the LA with 80% sensitivity, 80% specificity, and area under the ROC curve = 0.85 (95% confidence interval, 0.79-0.90) (Figure 3B).

Figure 3. A: impedance gradient between the pulmonary vein ostia (PVO) (by vein) and the left atrium (LA) in the first 15 patients. B: ROC curve of impedance to differentiate between LA and PVO; AUC, area under the curve.

In the final 10 patients an observer, who was only aware of the RB, classified as LA (≤RLA + 4% W) 206 (85%) of 242 CARTO points in the LA, and as PVO (>RLA LA + 4% W) 98 (83%) of 118 CARTO points in the PVO (91% positive and 73% negative predictive value for identifying the LA). Using the impedance lines as a guide, all the PV were isolated, and application in the interior of the crown was necessary in up to 10% of the PV. If the RB rose more than 4% W, the radiofrequency application was stopped. The mean radiofrequency time for the left circlet was 22 (9) min and for the right, 19 (6) min. There were no complications. Seventy percent of the patients were free of atrial arrhythmias at 5 (3) months. No PV stenosis was observed on computerized tomography of the chest.

DISCUSSION

A high to low linear impedance gradient exists from the deep PV to the LA, which is homogeneous and constant for each patient. It is possible to reconstruct a 3-dimensional impedance map of the LA and of the PV in which, adjusting the scale by 4% of the impedance above the mean value of the RLA, the transition zone can be defined so that circumferential isolation may be safely performed.

The baseline cardiac impedance value is determined by a multitude of factors, some biological, such as the volume of blood inside a chamber, its radius or the pulmonary resistivity adjacent to a vein.4-7 Extrinsic factors must also be kept in mind, such as the size and positioning of the reference patch, the size of the mapping catheter electrode, the pressure exerted on the tissue, the source of energy used and factors related to the patient, such as body surface.4-7 All this means that the RB varies for each investigator and each center, which is why values should be known by each. What is important is that, although each patient has a different RB, a linear and constant downward gradient is maintained from the interior of each PV to the LA.

The PV-LA junction is a histologically complex zone with myocardial fibers interwoven in various directions.8 Electrical resistivity is determined by the direction of the fibers9 and the degree of fibrosis, both of which can explain the differences in RB obtained in this zone. We prospectively validated that RLA + 4% W defines a safety zone in the LA to guide circumferential isolation. Unlike the model used by Lang et al,7 we consider a percentage and not an absolute increase in RLA to be more exact, as it varies from patient to patient.

Mapping of each PV, instead of the virtual tubes that are generally used, allows delimitation of their shape and observation of the impedance gradient when the catheter is advanced towards the PVO. In this way, small branches of the PV are identified and coded with higher impedance, thereby avoiding the risk of applying radiofrequency in the interior. It is also useful when there exist right accessory branches (2 in our population).

The main limitation of the study is the identification of the PVO in the absence of PV imaging.10 Imaging techniques, with their limitations, could have helped with a more exact localization of the PVO, but the combination we used of fluoroscopy, electrical signals and catheter movement can be recognized as valid. This study was limited to patients who underwent a single ablation procedure and who had paroxysmal atrial fibrillation. The results should not be translated to patients having already undergone ablation or with chronic atrial fibrillation.

In conclusion, 3-dimensional impedance mapping allows identification of the transition zone between the PVO and the LA, and is helpful for guiding circumferential isolation of the PV.

Adriano Jiménez-Velasco is the holder of a grant from Biosense-Webster.

Correspondence: Dr. A. Pedrote.
Unidad de Arritmias. Servicio de Cardiología. Hospital Virgen del Rocío. Manuel Siurot, s/n. 41013 Sevilla. España.
E-mail: pedroteal@hotmail.com

Received February 4, 2008.
Accepted for publication May 14, 2008.