Introduction

Pulmonary vein isolation (PVI) is a cornerstone of atrial fibrillation (AF) therapy. It consists in blocking the conduction of stimuli from the pulmonary veins (PVs) to the atrial tissue, thereby eliminating one of major AF triggers.1 As a complex procedure, PVI is undergoing continuous refinement aimed at improving its efficiency, simplicity, safety, and long-term efficacy.

Ablation technologies and equipment have been revised and upgraded over the last years. Contact force (CF) catheters and multihole irrigated tip technologies have been proven to enhance PVI effectiveness, and have been subsequently widely adapted.2-4 Moreover, emerging ablation technologies, such as pulsed-field ablation, expandable lattice-tip, or ultralow temperature cryoablation await verification of their promising potential.5

On the other hand, strategies and techniques using the available means have evolved along with the devices. Various modifications to the conventional PVI protocol have been proposed. Researchers found that touch-up ablations and low density of lesions should be avoided.6 Many adjunctive ablation strategies were also evaluated with some studies advocating7 and others undermining8,9 the effectiveness of this approach. However, available evidence suggests that creating additional lesions might reduce the AF recurrence rate.10

Despite the progress made, the success rate of PVI remains unsatisfactory. The reported reconnection rates reach over 70%.1 In available retrospective analyses, long-term freedom from recurrence is achieved in around 50% of patients.11,12 Data regarding the influence of the distance between ablation lines (DBL) on the success of the procedure are limited.

In this study, we aimed to determine the factors related to the procedural technique that impact the arrhythmia recurrence rate after PVI performed for paroxysmal or persistent AF.

Patients and methods

Study design

This retrospective cohort study was conducted at a tertiary care center in Poland, where approximately 600 procedures per year are performed, including catheter ablations of different atrial and ventricular arrhythmias, as well as electrophysiologic studies. In our center, PVI is the first procedure attempted in patients with AF, without additional linear / regional ablation (in the case of no other registered atrial arrhythmias), which is in line with the European Society of Cardiology guidelines.1 A total of 146 consecutive patients were enrolled, based on the following inclusion criteria: age of at least 18 years, undergoing a first-time PVI for either paroxysmal or persistent AF, and use of the CARTO 3D electroanatomic system (Biosense Webster, Inc., Irvine, California, United States) during PVI. The patients who underwent single-shot procedures, procedures performed with the use of a system other than CARTO 3D, or those with a history of surgical or catheter ablation for AF were excluded from the study. Two different ablation techniques were used in the study population: 75 patients underwent PVI with the use of a very-high-power, short-duration (vHPSD) catheter, QDot Micro (Biosense Webster, Inc.), while the remaining 71 patients were treated with an ablation index (AI)-guided ThermoCool Smarttouch Surround Flow (SF) catheter (Biosense Webster, Inc.). The choice of the method was made at the discretion of a physician prior to the procedure. Study procedures were performed between December 2019 and December 2021. The study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki. Its protocol was approved by the local Bioethics Committee (AKBE/127/2022). All patients gave their written informed consent to the processing of personal data.

Procedural workflow

Every PVI procedure was preceded with transesophageal echocardiography to rule out an intracardiac thrombus and to assess the anatomy of the interatrial septum. All catheters were inserted under local anesthesia through femoral vein punctures. During the procedure, unfractionated heparin was infused according to the activated coagulation time (target >335 s); the first bolus dose (100 IU/kg) was administered before transseptal puncture. Opioids (mostly continuous remifentanil infusion) were used for intraprocedural pain management. Some patients required sedation with midazolam boluses administered at the discretion of an operator. In most procedures, a 3-dimensional reconstruction of the left atrium (LA) and PVs was rendered using rotational angiography. Bipolar voltage mapping was generated using either a 20-pole Nav Lasso or a PentaRayTM catheter (Biosense Webster, Inc.) along with the CARTO 3D electroanatomic navigating system (Biosense Webster, Inc.). The QMode+ algorithm (temperature-controlled ablation) was used to perform vHPSD ablation (90 W, 4 s) with a QDot Micro catheter. Each radiofrequency (RF) application in this mode was preceded by 2 seconds of precooling and 4 seconds of irrigation flow (8 ml/min). The operators kept a distance of up to 4.5 mm between lesions on the anterior wall, and 5 mm in other regions. A temperature of 55 °C, based on the thermocouple with the highest temperature, was set as a cutoff. AI-guided procedures with the use of a Thermocool Smarttouch SF catheter were performed according to the CLOSE protocol. DBLs were measured with the CARTO 3D “line” tool. If the distance was greater than predefined in the protocol, additional lesions were created, even when isolation was proved. Power output was 35 W, with a target AI of over 400 at the posterior and inferior LA walls and over 550 at the remaining sites. Other settings included target CF of 10 to 30 g, irrigation rate of 15 ml/min, a maximum interlesion distance of 6 mm, and maximum temperature cutoff of 40 °C. Procedural success was confirmed once entrance block to all PVs was proven. In the case of any doubts, exit block was also checked by pacing from the PVs. Assessment of entrance block was performed with a diagnostic multipolar catheter placed separately in every PV, with a reassessment 15 minutes after the last application. The block was considered a final proof of the acute durability of linear lines. Additional RF applications were delivered if short-term PV reconnection occurred. Pericardial effusion and other intracardiac complications were excluded with transthoracic echocardiography performed twice: immediately after the procedure and in the morning on the following day. All PVIs were carried out by experienced operators who perform over 50 of such procedures per year.

Measurement methods

In all patients, fast anatomical maps (FAMs) created before the procedure were analyzed, and the minimal distance between opposite PVI lines was measured. DBL was defined as the minimal distance between 2 VisiTag locations with targeted AI values, connected to the isolation lines at the opposite sites. In a majority of the patients, DBL was measured in the roof of the LA or in the superior part of the LA posterior wall. Transverse diameter (TD) of the LA was defined as the distance between PV carinas at the opposite sites. Assuming a greater DBL in a patient with a greater TD, the DBL/TD ratio was included in the analysis. Moreover, the posterior-anterior distance was measured as a maximum distance between posterior and anterior LA walls in the transverse plane. In every patient, the initial impedance level of VisiTag points in the uppermost, lowermost, posterior carina, and anterior carina sites of the ipsilateral PVI line were calculated, along with the mean value from these 4 assessments.

Study end points

The primary end point was freedom from atrial arrhythmia (AF, atrial flutter [AFl], or atrial tachycardia [AT]) 12 months after the procedure. Influence of DBL and the DBL/TD ratio on arrhythmia recurrence was analyzed in the whole group, and in the vHPSD and AI-guided ablation subgroups.

Follow-up of the patients

The data on end points were extracted from patient medical records. Clinical follow-up data were gathered during routine postdischarge appointments at the outpatient clinic. Each patient had 24-hour Holter monitoring scheduled at 3 and 12 months after the ablation. Then, the patients were contacted by telephone for additional information, including any postablation symptoms of arrhythmia, documentation of potential arrhythmia recurrence, and the precise time of its occurrence. Recurrence of arrhythmia was defined as an episode of AF, AFl, or AT lasting at least 30 seconds recorded on a Holter or 12-lead electrocardiography. First 3 months after the procedure were considered the blanking period.

Statistical analysis

Distribution of continuous variables was assessed with the Shapiro–Wilk test. Due to a non-normal distribution, the results are presented as median and interquartile range (IQR). Categorical variables are presented as percentages. The Fisher exact test was used to compare categorical variables, and the Mann–Whitney test was used for continuous variables. To assess the risk of arrhythmia recurrence, univariable and multivariable stepwise Cox proportional hazards regressions were performed, assuming a P-entry value of 0.2 and a P-stay value of 0.1. To assess the influence of the measured parameters on the impedance level, Spearman correlation coefficients were calculated. The Kaplan–Meyer survival curves were plotted for an analysis of the arrhythmia recurrence. The log-rank test was performed and its results are visualized in the Figures showing the Kaplan–Meyer survival curves. The cutoff point for the DBL/TD ratio was determined with the Youden index method, and an area under the receiver operating characteristic (ROC) curve (AUC) for this value was calculated. A P value below 0.05 was considered significant. The statistical analysis was performed using Statistical Analysis Software, version 9.4 (Cary, North Carolina, United States).

Results

Patient characteristics

A total of 146 patients were included in our study. The median (IQR) patient age was 62 (53–68) years, 34.3% of the participants were women, and the median body mass index (BMI) was 28.1 (25.2–31.2) kg/m2. A majority of the population had hypertension (61%), while only every tenth patient (11%) was diabetic. Paroxysmal AF was over 2-fold more prevalent than persistent AF (70.6% vs 29.5%). More than 80% of the study population were treated with a β-blocker, whereas only a small group of patients had other antiarrhythmics prescribed (amiodarone, 2.7%; propafenone, 4.1%). The abovementioned variables showed no significant differences between the vHPSD and AI-guided groups. Baseline characteristics of the whole population and of individual subgroups are listed in Table 1. The median (IQR) follow-up duration was 33 (12–53) weeks.

Table 1. Baseline characteristics of the study population

Variable

Total (n = 146)

vHPSD (n = 75)

AI-guided (n = 71)

P value (vHPSD vs AI-guided)

Demographic and anthropometric data

Age, y

62 (53–68)

63 (55–69)

60 (49–68)

0.27

Women

34.3

38.7

29.6

0.3

Weight, kg

84 (75.5–95.5)

83 (73–95)

86 (77–97)

0.5

Height, cm

175 (166–181)

174 (164–182)

175 (167–180)

0.81

BMI, kg/m2

28.05 (25.15–31.18)

28.05 (24.77–31.26)

27.61 (25.31–30.88)

0.99

Comorbidities

Hypertension

61

61.3

60.6

>0.99

DM

11

10.7

11.3

>0.99

AF type

Paroxysmal

70.6

73.3

67.6

0.47

Persistent

29.5

26.7

32.4

Medications after the procedure

Amiodarone

2.7

4

1.4

0.62

Propafenone

4.1

5.3

2.8

0.68

β-Blocker

81.5

77.3

85.9

0.21

Data are presented as median (interquartile range) or percentage of patients.

Abbreviations: AF, atrial fibrillation; AI, ablation index; BMI, body mass index; DM, diabetes mellitus; vHPSD, very-high-power, short-duration ablation

Procedural characteristics

A total of 102 PVI procedures (69.9%) performed in the study population started with a sinus rhythm (SR). Average LA dimensions were 37.05 mm and 76.8 mm for the posteroanterior and transverse diameters, respectively (based on FAM). Median (IQR) duration of the procedure was 135 (110–160) minutes, while median (IQR) fluoroscopy time reached 422.5 (299–629) seconds. A median (IQR) of 75 (61–90) RF applications per procedure were performed. DBL was similar between the subgroups, with median (IQR) values of 35.9 (29.6–43.5) mm for the vHPSD and 33.7 (27.3–40.1) mm for the AI-guided group. There was no significant difference in the DBL/TD ratio (Figure 1). Only 3 out of 10 procedural variables (mean right-sided impedance, mean left-sided impedance, and procedure time) were found to differ significantly depending on the ablation technique used. Procedural data are presented in Table 2.

Figure 1. Box plot for the ratio of distance between the opposite pulmonary vein isolation lines (DBL) to the transverse diameter (TD) of the left atrium in the groups with and without atrial fibrillation (AF) recurrence

Table 2. Procedural data, study end points, and safety analysis

Variable

Total (n = 146)

vHPSD (n = 75)

AI-guided (n = 71)

P value (vHPSD vs AI-guided)

SR at the start of the procedure

69.9

70.7

69

0.86

PAD, mm

37.05 (31.6–42.5)

37.1 (32.6–41.2)

36.8 (30.5–42.9)

0.71

TD, mm

76.8 (70.1–84.1)

78.5 (69–85.4)

75.3 (70.1–82.8)

0.35

DBL, mm

34.65 (28.4–41.2)

35.9 (29.6–43.5)

33.7 (27.3–40.1)

0.12

DBL/TD ratio

0.47 (0.4–0.52)

0.48 (0.41–0.55)

0.45 (0.37–0.52)

0.17

Mean right-sided impedance, Ohms

127.25 (111.5–136.13)

108.88 (103.75–115.75)

134.25 (129–140.75)

<⁠0.001

Mean left-sided impedance, Ohms

122.63 (111.25–133.38)

108.50 (105.25–115.25)

131.00 (123.50–138.25)

<⁠0.001

Procedure time, min

135 (110–160)

120 (95–140)

145 (130–165)

<⁠0.001

Fluoroscopy time, s

422.5 (299–629)

393 (248–629)

439 (339–660)

0.15

Number of RF applications

75 (61–90)

76 (61–89)

74 (60–94)

>0.99

AF recurrence (overall)

33.6

33.3

33.8

>0.99

AF recurrence before discharge

4.1

4.1

4.2

>0.99

Postprocedural pericardial effusion

0

0

0

–

Vascular complications

3.4

4

2.8

>0.99

Pseudoaneurysm

0.7

1.3

0

>0.99

Femoral arteriovenous fistula

1.4

2.7

0

0.5

Cerebrovascular event

0.7

1.3

0

>0.99

Data are presented as median (interquartile range) or percentage of patients.

Abbreviations: DBL, distance between opposite pulmonary vein isolation lines; PAD, posteroanterior diameter of the left atrium; RF, radiofrequency; SR, sinus rhythm; TD, transverse diameter of the left atrium; others, see Table 1

Clinical outcomes

Recurrence of atrial arrhythmia was observed in 49 patients (33.6%). In this group, the most common form of atrial arrhythmia was AF (n = 44; 89.8%), followed by AFl (n = 4; 8.2%) and AT (n = 1; 2%). In 3 of the 4 patients with AFl recurrence, the arrhythmia was dependent on the LA posterior wall. Treatment-emergent adverse events occurred at a rate of less than 4%, irrespective of the ablation technique used (Table 2).

Predictors of atrial fibrillation recurrence

Median (IQR) values of DBL and the DBL/TD ratio were 40.4 (36.6–45.2) mm and 0.48 (0.47–0.49), respectively. In the univariable regression analysis (Table 3), both greater DBL and a greater DBL/TD ratio were associated with a reduced risk of AF recurrence in the whole population (hazard ratio [HR], 0.966; 95% CI, 0.935–0.998 [per 1 mm]; P = 0.04 and HR, 0.968; 95% CI, 0.944–0.993 [per 1%]; P = 0.01, respectively). Such a correlation was not observed in the vHPSD group, while in the AI-guided group only the DBL/TD ratio was found to be associated with AF recurrence (HR, 0.949; 95% CI, 0.914–0.986; P = 0.007). No other variable was shown to positively or negatively correlate with AF recurrence (Table 3). In the whole population, the optimal cutoff value for the DBL/TD ratio was 45.7%. In the ROC analysis, this DBL/TD value showed moderate strength for the prediction of AF recurrence (AUC = 0.641) (Figure 2). Estimated changes in the AF recurrence rate in 2 groups stratified based on the DBL/TD cutoff value (P = 0.01) are presented in Figure 3. The correlation analysis revealed no difference in impedance levels regardless of the DBL/TD values (right side, P = 0.18; left side, P = 0.38) (Figure 4).

Table 3. Results of the regression analysis for time to arrhythmia recurrence

Variable

HR

95% CI

P value

Total (n = 146)

DBL, per 1 mm

0.966

(0.935–0.998)

0.04

PAD, per 1 mm

1.031

(0.995–1.068)

0.09

TD, per 1 mm

1.01

(0.98–1.04)

0.53

DBL/TD ratio, per 1%

0.968

(0.944–0.993)

0.01

Mean right-sided impedance, per 1 Ohm

0.999

(0.98–1.018)

0.9

Mean left-sided impedance, per 1 Ohm

1.003

(0.982–1.025)

0.75

Number of RF applications

1.005

(0.995–1.016)

0.33

vHPSD (n = 75)

DBL, per 1 mm

0.978

(0.938–1.02)

0.31

PAD, per 1 mm

1.029

(0.977–1.084)

0.29

TD, per 1 mm

0.991

(0.952–1.033)

0.68

DBL/TD ratio, per 1%

0.245

(0.007–9.136)

0.45

Mean right-sided impedance, per 1 Ohm

0.981

(0.933–1.031)

0.45

Mean left-sided impedance, per 1 Ohm

0.983

(0.931–1.038)

0.54

Number of RF applications

1.001

(0.985–1.017)

0.91

AI-guided (n = 71)

DBL, per 1 mm

0.95

(0.902–1.001)

0.05

PAD, per 1 mm

1.034

(0.985–1.086)

0.18

TD, per 1 mm

1.029

(0.986–1.073)

0.19

DBL/TD ratio, per 1%

0.005

(0–0.246)

0.007

Mean right-sided impedance, per 1 Ohm

0.983

(0.946–1.022)

0.39

Mean left-sided impedance, per 1 Ohm

0.999

(0.962–1.037)

0.96

Number of RF applications

1.009

(0.994–1.025)

0.24

Abbreviations: HR, hazard ratio; others, see Tables 1 and 2

Figure 2. Receiver operating characteristic curve with area under the curve (AUC) analysis for the ratio of the distance between opposite pulmonary vein isolation lines to the transverse diameter of the left atrium

Figure 3. Kaplan–Meier analysis of freedom from atrial fibrillation (AF) recurrence in patients stratified based on the ratio of the distance between opposite pulmonary vein isolation lines (DBL) to the transverse diameter of the left atrium (TD)

Figure 4. Scatter plot for impedance levels based on the ratio of the distance between opposite pulmonary vein (PV) isolation lines (DBL) to the transverse diameter (TD) of the left atrium

In the multivariable analysis, only the DBL/TD ratio and a history of stroke were identified as independent risk factors for AF recurrence (HR, 0.969; 95% CI, 0.942–0.996 [per 1%]; P = 0.03 and HR, 14.3; 95% CI, 1.712–120.2; P = 0.01, respectively).

Discussion

The major findings of our analysis are as follows: 1) the greater the DBL and its ratio to the TD, the lower the risk of AF recurrence in the population of 146 patients undergoing the first PVI procedure (every 1-mm increase in DBL reduced the risk of AF recurrence by about 3%); 2) a lack of a correlation between the DBL/TD ratio and impedance levels proves that ablation sites were localized in the atrial parts of the PV ostia.

Catheter ablation for PVI intended to restore and / or maintain SR is recommended in AF patients when drug therapy fails or, in some cases, as a first-line treatment.1 As an invasive method of AF management, PVI exposes a patient to additional risk of procedural complications, which should be compensated by a long-lasting therapeutic effect; however, it is often not the case, as the post-PVI arrhythmia recurrence rate is still relatively high. Nonetheless, in this respect, catheter ablation is superior to antiarrhythmic drug therapy, as it is associated with a higher freedom from arrythmia recurrence and greater improvement in quality of life, without an increase in the rate of serious adverse events.13 Since all kinds of improvement are highly desirable, procedural modifications that would lead to a higher success rate are being continuously sought after. We found that increasing the DBL and DBL/TD ratio during PVI is beneficial. This is a modification of the strategy rather than of technology; the proposed change in the definition of optimal lesion localization does not require additional equipment or extensive training, and is easy to adapt in the PVI protocol in every electrophysiology laboratory performing PVI (Figure 5). Similar revisions of what currently is the most common approach have already been debated in the literature. Among the suggested solutions we can find addition of 6 short ablation lines on the PVI circumferences,10 posterior wall isolation (PWI),14,15 low-voltage area (LVA) ablation,16 adjunctive electrogram-guided ablation,17 interpulmonary isthmus ablation,9 or simply focusing on sufficient density and CF, with no touch-up lesions.6 Even though all the abovementioned PVI modifications or adjunctive methods have some theoretical justification, the results of studies evaluating these approaches varied considerably, with some strategies (6 additional lines, stepwise ablation) being supported by evidence of improvement in the freedom from AF recurrence, while other (interpulmonary isthmus or PWI ablation) showing no effects on this end point. The approach involving an increase in DBL and the DBL/TD ratio proposed in this study has a pathophysiologic basis. More distal RF application is associated with a greater chance to achieve transmurality, as the atrial tissue in this location is thinner, with a lower number of epicardiac muscle fibers. Moreover, the circuit of lines required for complete isolation is smaller, making this approach less demanding in technical terms. In contrast to PVI modifications or adjuncts with no proven efficiency, the accuracy of our suggestion is reflected by results of the AUC analysis showing significant strength of prediction. Interestingly, in a study by Krzyżanowski et al,18 DBL showed no impact on AF recurrence in the long-term follow-up of first-pass isolation. The contradiction between our findings may arise from methodologic differences. Firstly, in the mentioned study, DBL was measured between the uppermost and the lowest points of 2 opposite isolation lines, while in our study it was measured between the nearest points (Figure 5). Moreover, and perhaps more importantly, Krzyżanowski et al18 did not index the DBL to the total LA diameter. Therefore, in the case of enlarged atria, even with high DBL values, isolation lines might be far from the PVs. Nevertheless, participants of the 2 studies seem to share some important baseline characteristics: similar age, sex distribution, paroxysmal AF share, and BMI, which is especially important since higher BMI values were proven to be associated with shorter freedom from AF recurrence.19,20 In view of the above, further studies investigating the role of DBL and the DBL/TD ratio in a larger population are required to confirm their importance and efficacy of their enhancement.

Figure 5. Comparison of a short (A) and long (B) distance between ablation lines after a pulmonary vein isolation procedure. Red dots mark the radiofrequency application spots, red arrows mark the left side of the patient (L), green arrows mark the right side (R).

Further investigations are also needed to assess the safety of the proposed PVI optimalization strategy. In this study, we did not compare the rate of complications between the groups with wide and narrow DBL, or between individuals with a significant difference in the DBL/TD ratio. In this regard, PV stenosis (PVS) appears to deserve special attention. PVS complicates 1% to 3% of catheter AF ablations in experienced centers; however, if lesions are applied relatively deeply, the rate may rise to 42.4%.21,22 This potentially morbid condition may vary in severity, with only a small percentage of cases necessitating intervention, so the potential rise in the PVS rate would not automatically mean the technique is contraindicated,23 but a close clinical assessment would be recommended. We expect that an optimal DBL range and / or DBL/TD range could be proposed to keep those parameters sufficiently high, but not too high, so as to avoid PVS. Importantly, we did not observe a correlation between DBL and the impedance level, which proves that ablation lines were not created in PVs.

Patients treated with 2 different ablation techniques, vHPSD or AI-guided ablation, were included in this study. Differences between those 2 subgroups were consistent with those reported in the available literature in terms of procedure duration (shorter in the vHPSD group) and the rate of complications (similar in the 2 groups) (Table 2).4,24,25 Importantly, the ablation technique did not affect the DBL/TD ratio. However, a lack of an influence of the DBL/TD ratio on the AF recurrence rate in the vHPSD group may be due to a difference in biophysical heating conduction, as compared with AI-guided ablation. Considering LA wall thickness, both methods have some advantages and disadvantages. AI-guided lesions tend to be deeper, which may influence the quality of isolation; however, their depth can increase collateral tissue damage. vHPSD produces effective tissue injury while limiting catheter instability and tissue edema formation. It is achieved by means of a modification in the biophysical properties of the traditional RF lesion: the higher power extends the resistive heating phase, resulting in immediate and lethal heating of the full thickness of the atrial wall, while the shorter duration reduces the length of the convective heating phase, limiting collateral tissue damage and inadvertent catheter displacement.26,27

Based on the study results, a greater DBL is favorable in terms of arrhythmia recurrence. We recognize the potential risk of some proarrhytmic areas located on the level of PV ostia not being isolated from the LA tissue in the case of maximalized DBL. This was also a hypothesis important for the strategy of wide-area circumferential ablation. However, our analysis showed that this strategy may even increase the risk of arrhythmia recurrence. What is more, a shorter DBL may promote the incidence of atypical AFl.

Limitations

Our study has several limitations. Firstly, it was a retrospective cohort study, where the groups were not randomized, nor were the characteristics of the included patients controlled. Secondly, the analyzed parameters were not indexed to the LA dimensions other than TD, which could have enabled a better discrimination between the patients with maximized DBL vs the patients with a large LA. Echocardiographic data were lacking, as they were not included in the medical records. Oral anticoagulation therapy was not evaluated in the study population; however, it is usually suboptimally prescribed, especially in older patients, which could affect the rate of bleeding and ischemic complications.28 Lastly, our study was conducted in a medium-sized group susceptible to bias, and the results obtained should be verified on a larger sample.

Conclusions

A close distance between PVI lines may contribute to AF recurrence; thus, increasing DBL and ensuring a higher DBL/TD ratio may be an advantageous ablation strategy. Even in the case of a higher DBL/TD ratio, the ablation sites were localized in the atrial parts of the PV ostia.