OUP user menu

Radiofrequency ablation of drug-refractory atrial fibrillation: an observational study comparing ‘ablate and pace’ with pulmonary vein isolation

Alessandro Proclemer, Giuseppe Allocca, Dario Gregori, Carlo Bonanno, Renato Ometto, Alessandro Fontanelli, Roberto Mantovan, Martino Crosato, Vittorio Calzolari, Daisy Pavoni, Domenico Facchin, Luca Rebellato, Paolo M. Fioretti
DOI: http://dx.doi.org/10.1093/europace/eun197 1085-1090 First published online: 30 July 2008


Aims To compare clinical characteristics, procedure complexity, acute and long-term outcome of ‘ablate and pace’ (A&P) with pulmonary vein isolation (PVI) in patients with drug-refractory atrial fibrillation (AF). So far, only few small studies have compared the two procedures.

Methods and results We analysed retrospectively a cohort of symptomatic consecutive patients with drug-refractory AF. Group 1 included 100 patients treated with A&P and Group 2 included 144 patients treated with PVI. Group 1 patients were older (74 ± 8 vs. 56 ± 9 years; P < 0.0001), had lower left ventricular ejection fraction (50 ± 13% vs. 59 ± 7%; P < 0.05), and a lower prevalence of paroxysmal AF (46% vs. 65%; P < 0.05). Acute success was not statistically different (98% vs. 92.3%, P = ns). Group 1 patients had shorter procedure time and lower radiation exposure with respect to Group 2 patients (70 ± 15 vs. 204 ± 58 min, and 8 ± 4 vs. 57 ± 22 min; P < 0.0001, respectively). After a median follow-up of 29 months (I, III quartile; 15, 40 months) vs. 25 months (I, III quartile; 8, 36 months) (P = ns), all the patients in Group 1 were free of symptomatic AF, while 113 patients (79%) of Group 2 were in stable sinus rhythm (P < 0.0001). Persistent or permanent AF has been documented in 58 patients (58%) of Group 1 vs. 11 (8%) of Group 2 (P < 0.0001).

Conclusion In this series (i) patients treated with A&P and PVI for drug-refractory AF showed significant differences in clinical profile; (ii) A&P is a shorter and less complex procedure, but is associated with a higher rate of persistent AF; (iii) symptomatic recurrences of paroxysmal AF were more frequent in PVI group. Randomized studies appear necessary to identify the best strategy in selected cases.

  • Ablate and pace
  • Isolation of the pulmonary veins
  • Drug refractory atrial fibrillation


Radiofrequency (RF) ablation of the atrioventricular junction (AVJ) plus pacing therapy—‘ablate and pace’ (A&P)—and electrical isolation of pulmonary veins (PVI) are effective non-pharmacological therapeutic strategies in patients with medically refractory atrial fibrillation (AF). A&P is a palliative therapy which improves important cardiac symptoms, quality-of-life indexes, and left ventricular (LV) function; consequently has been extensively utilized, also because of its relatively low procedural risk.17 PVI, instead, is a curative therapy for both paroxysmal and persistent AF with a non-negligible rate of AF recurrence and potential complications, such as stenosis of pulmonary veins (PVs), stroke, and cardiac tamponade.814

To the best of our knowledge only one published study compared these procedures in a small series of elderly patients.15 The importance of a comprehensive evaluation of non-pharmacological rate- or rhythm-control strategies, such as A&P and PVI has recently been stressed by a task-force committee of the AHA/ACC/ESC.16 Therefore, we conducted an observational study comparing clinical characteristics, methodological complexity, acute and long-term outcome, and major adverse effects of two historical groups of consecutive patients, one treated with A&P and one with PVI. The correct identification in the single case of the best ablation strategy appears fundamental to obtain good clinical effects with lowest risk.16,17


From January 2003 to December 2006, 244 consecutive patients with drug-refractory AF were referred to three Electrophysiology Laboratories of north-eastern Italy for non-pharmacological treatment of AF. The choice between A&P and PVI was based upon the clinical characteristics of each case and after having discussed with the patients the potential benefit and possible risks of the two procedures. The data used for the study was entered prospectively into a common database.

The three hospitals had greater than 10 years experience in the field of A&P therapy, while their experience with PVI begun in 2002. All patients gave written informed consent for catheter ablation of AVJ or PVI and, in case of A&P therapy, for pacemaker implant. The study population consists of 100 consecutive patients (Group 1) who underwent A&P and 144 consecutive patients (Group 2) who were treated with PVI for drug-refractory paroxysmal, persistent, or permanent AF.

According to guidelines,16 AF is designated paroxysmal if the arrhythmia terminated spontaneously, and persistent when sustained beyond 7 days. Termination with pharmacological therapy or direct-current cardioversion does not alter the designation. The category of permanent AF included cases of long-standing AF and cases in which cardioversion has failed or has been foregone.16

The baseline and outcomes data compared between Groups 1 and 2 included: demographics, AF type, left ventricular ejection fraction (LVEF), procedural data (acute success, fluoroscopy and procedure times, possible complications), mortality, cause of death, New York Heart Association (NYHA) class at enrolment and end of follow-up, cerebral vascular accidents, symptomatic and asymptomatic AF recurrences, appearance of persistent or permanent AF in patients with paroxysmal AF.

Ablate and pace therapy

The AVJ ablation was obtained by the techniques previously described.4,5 Briefly, RF energy was delivered between the distal 4-mm electrode of a 7 Fr ablation catheter and dispersive patch-electrode. The goal of the procedure was the induction of complete AV block (acute success). In the majority of patients, the pacemaker was implanted immediately before AVJ ablation.4,13 In this case, the pacemaker was temporarily programmed to VVI mode at 30 bpm, and a safe distance between permanent leads and ablation catheter was fluoroscopically monitored during RF energy delivery. Dual-chamber rate responsive units with mode switch algorithms were used in patients with paroxysmal AF while single-chamber rate-adaptive systems were selected for patients with permanent AF.

Catheter isolation of pulmonary veins isolation

The electrophysiologically-guided ablation technique used has been previously described by other authors11,18,19 and by us.20 The left atrium was explored by either single or dual transseptal technique after confirmation of the absence of intracardiac thrombi by transoesophageal echocardiography. Following transseptal access, heparin was administered to maintain an activaction clotting time of approximately 250 s. Mapping of PVs was performed by means of a circumferential multipolar steerable catheter (LASSO, Biosense Webster, Diamond Bar, CA, USA) inserted into all PVs and positioned orthogonal to the PV axis as proximal as possible to the ostium. A 4 mm-tip catheter in the first 10 patients or a 3.5 irrigated tip catheter for mapping and ablation in the remaining 134 patients (Celsius Thermocool or Navistar, Biosense Webster) was introduced into left atrium and stabilized with the aid of a long sheath at the left atrium-PV junction. The ostium of PVs was defined by the following characteristics: (i) the angiographically defined ostium; (ii) the point where the catheter tip suddenly dropped into the atrium; (iii) the point where the PV potential and the atrial potential were the closest. In 76 patients, a non-fluoroscopic navigation system (CARTO, Biosense Webster) was integrated to construct three-dimensional electroanatomical maps of the left atrium and PVs.20 RF pulses around the single ostia of PVs were applied for 30–60 s with a power limit of 40 W and temperature limit of 45°. The endpoint was the electrical disconnection of the right and left PVs (acute success). Cavotricuspid isthmus ablation was performed et the end of the procedure in patients with documented episodes of common atrial flutter.

Post-procedure management and follow-up

In Group 1, clinical follow-up and pacemaker checks were obtained at 6-month interval. All patients continued oral anticoagulants. Main outcome parameters were analysed at each follow-up visit. A second AVJ ablation was effected in the case of AVJ conduction relapse. LVEF was re-assessed at the last follow-up visit.

In Group 2, all the patients were discharged on oral anticoagulants, and continued antiarrhythmic drugs usually at lower doses for 3 months. Clinical examination and 24-h Holter monitoring were performed after 1, 3, 6, 12 months, then every 6 months. Antiarrhythmic drugs were continued after the first 3 months in patients with AF recurrences. At each follow-up visit, the same clinical endpoints considered in Group 1 have been analysed. Freedom from recurrent AF was defined as the absence of symptomatic AF and asymptomatic sustained AF (lasting more than 30 s) recorded by Holter monitoring or ECG. The first month after PVI procedure was excluded from the analysis. A second AF ablation procedure was performed in case of frequent and symptomatic AF recurrences. Kaplan–Meier analysis was obtained for symptomatic AF recurrence in Group 1 vs. Group 2 patients.

Statistical analysis

Distribution of baseline characteristics and study outcomes between A&P and PVI groups has been described using median (interquartile ranges) or percentages (number of observations) whenever appropriate. Univariable analysis, based on unadjusted tests of significance between groups, both for baseline characteristics and for outcomes, has been based on Wilcoxon test for continuous variables and on Fisher exact test for categorical ones.

To properly account for baseline unbalance in Group 1 and Group 2 with respect to baseline characteristics, effects of A&P and PVI on outcomes data have been evaluated adjusting for gender, age, LVEF, NYHA class, documented underlying heart disease, and AF type (paroxysmal vs. persistent/permanent). According to the type of outcome, logistic regression (probability of acute success, incidence of major complications) or Cox proportional hazard model (procedure time, fluoroscopy time, time to death) was used. In addition, to obtain a more conservative estimate of the cohort effect, represented by the different periods of enrolment of the A&P and PVI groups and in view of the possible unobserved variables, which could potentially influence the baseline characteristics (both observed and not observed) of the patients, an individual, per-patient frailty was introduced. Then, a shared frailty was introduced to accommodate for unobserved centre characteristics, which could have also influenced the baseline survival, such as a hidden selection bias or organizational procedures influencing the outcome. Frailties' significance was tested using an approximate Wald test. A gamma distribution was assumed for the frailty and its sensitivity to the prior distribution chosen was assessed comparing the results with a Gaussian prior. All P-values, unless explicitly stated, refer to a 0.05 threshold. All analyses were performed in R system (R Foundation for Statistical Computing).21,22


Clinical characteristics

The baseline characteristics of both the groups are summarized in Table 1. Group 1 was represented with respect to Group 2 by older patients (74 ± 8 vs. 56 ± 9 years; P < 0.001) with higher prevalence of organic heart disease (74% vs. 41%; P < 0.001), NYHA class III–IV (56% vs. 6%; P < 0.001), lower mean value of LVEF (50 ± 13% vs. 59 ± 7%; P < 0.05), and lower prevalence of paroxysmal AF (46% vs. 65%; P < 0.05). The number of previous ineffective antiarrhythmic drugs was similar between the two groups. In Group 1, 49 patients received a single-chamber VVI-R permanent pacemaker, 47 patients a dual-chamber DDD-R pacemaker, and only four patients a CRT-P pacemaker for biventricular pacing.

View this table:
Table 1

Baseline clinical characteristics

Procedural findings

Acute success of both the procedures was not significantly different (98 vs. 92.3%, P = ns) (Tables 2 and 3); in particular, the induction of third-degree AV block has been obtained in 98 patients of Group 1 and a mean of 3.8 ± 0.4 PVs per patient have been disconnected in Group 2. Total procedure, fluoroscopy, and RF delivery times were significantly higher in Group 2 (Table 2). After adjustment for observed factors of unbalancing and unobserved frailty effects, procedure time and fluoroscopy time were significantly shorter in the A&P group than in the PVI group (respectively, −127 min; 95% CI: −146.95, −108.14; and −44.27 min; 95% CI: −57.15, −31.38) (Table 3). Risk of complications was also significantly lower for A&P than for PVI (OR 0.02, 95% CI 0.001–0.62) (Table 3). In Group 1, one patient was promptly cardioverted few hours after A&P because of ventricular fibrillation, and one patient needed pacemaker replacement after 2 weeks because of early battery dysfunction. In Group 2, five patients undergoing the PVI procedure developed cardiac tamponade effectively treated with percutaneous pericardial drainage, one patient reported mitral valve entrapment of LASSO catheter that was resolved by gentle traction of the catheter. No cerebral complications have been observed in both the groups during the procedures and in the following 30 days.

View this table:
Table 2

Procedural findings

View this table:
Table 3

Effects on outcome by multivariate analysis


All the patients were followed-up for at least 6 months with a median follow-up of 29 months (I, III quartile; 15, 40 months) in Group 1 and 25 months (I, III quartile; 8, 36 months) in Group 2 (P = ns). After ‘A&P’ 98 patients (98%) were free of symptoms related to AF recurrence, despite persistent or permanent AF were documented by ECG and atrial rate histograms of pacemaker diagnostic function in 58 of them (58%). Out of 56 patients in NYHA class III–IV before A&P, 46 patients showed a persistent improvement of the functional class (I–II) after the procedure. The mean LVEF value increased from 50 ± 13% to 54 ± 12% (P = 0.05).

In the group of patients treated by PVI, 113 (79%) were in stable sinus rhythm, 20 patients (14%) presented paroxysmal AF recurrence, and 11 (8%) showed persistent or permanent AF. The group of patients without AF recurrences showed, in comparison with the group of patients with symptomatic and asymptomatic AF recurrences, similar mean age (56 ± 9 vs. 55 ± 9, P = ns), male prevalence (74 vs. 77%, P = ns), documented underlying heart disease (41 vs. 42%, P = ns), NYHA functional class III–IV (4 vs. 10%, P = ns), mean LVEF value (59 ± 8 vs. 59 ± 6%, P = ns), mean fluoroscopy and RF delivery times (56 ± 20 vs. 59 ± 30 min, and 35 ± 11 vs. 32 ± 12 min, respectively, P = ns), but higher prevalence of paroxysmal AF (72 vs. 39%, P < 0.01), higher acute success rate (97 vs. 77%, P < 0.01), and lower mean procedure time (199 ± 53 vs. 225 ± 74 min, P < 0.01) (Tables 3 and 4).

View this table:
Table 4

Long-term results

By Kaplan–Meier analysis, 98% of patients in the ‘A&P’ group and 70% of those in the PVI group were free from AF-related symptoms (P < 0.05) (Figure 1). A second procedure was necessary owing to symptomatic AF recurrences in two patients (2%) of Group 1 and in 20 patients (14%) of Group 2 (P < 0.01). At the last follow-up visit, persistent use of antiarrhythmic drugs was significantly lower in Group 1 than in Group 2 (9 vs. 48%; P < 0.0001). In particular, nine patients in Group 1 were on sotalol therapy, 30 patients in Group 2 were treated by sotalol, 20 by propafenone, and 20 by flecainide. Eighty-four patients of the Group 1 and 44 of the Group 2 were on persistent oral anticoagulant therapy (84 vs. 31%, P < 0.0001).

Figure 1

Freedom from symptomatic recurrent atrial fibrillation after A&P (Group 1) and pulmonary vein isolation (Group 2).

Sixteen patients (16%) in Group 1 and no patient in Group 2 died during the follow-up period. After A&P, 10 patients died due to non-cardiac causes, four owing to refractory heart failure, two of sudden death (4 and 24 months after the procedure, respectively). In Group 1, comparison between the 84 survivors and 16 non-survivors at the end of follow-up revealed similar mean age (73 ± 8 vs. 76 ± 8, P = ns), male prevalence (43 vs. 63%, P = ns), documented organic heart disease (73 vs. 81%, P = ns), history of paroxysmal AF (48 vs. 38%, P = ns), but higher mean LVEF value (51 ± 13 vs. 41 ± 13, P < 0.05), and lower prevalence of NYHA functional class III–IV (48 vs. 75%, P < 0.05). One 69-year-old patient in Group 2 suffered an irreversible metabolic coma, unrelated to PVI, 5 months after the procedure. After adjustment for unbalanced factors and for frailties, risk to die was not significantly higher for A&P than for PVI (Table 3).


At present, A&P is a well-established palliative therapy in patients with symptomatic drug-refractory paroxysmal or permanent AF, rapid ventricular rate, or tachycardia-induced cardiomyopathy despite some disadvantages, such as the persistent need of oral anticoagulation, life-long pacemaker dependency, loss of AV synchrony, and possible risk of sudden death.17 Meta-analysis of 21 randomized clinical trials demonstrated that AVJ ablation plus permanent pacemaker implantation significantly improves symptoms, quality of life, exercise capacity, and reduces hospital admissions.6 Based on randomized1,2 and observational studies,3,5 recent guidelines16 for the management of patients with AF included A&P in class IIA (level of evidence B) of recommendations for non-pharmacological ‘rate-control strategy’. In contrast, PVI by means of RF ablation represents an effective and curative therapy in approximately 70–80% of patients with drug-refractory paroxysmal AF and in approximately 60–70% of patients with persistent or permanent AF, independently of the methodology used.9,18,23,24 After the publication of observational studies,1013,19,25,26 large literature review,23 and multicentre experience,9 the abovementioned guidelines16 considered AF catheter ablation for maintenance of sinus rhythm a reasonable alternative to pharmacological therapy only in symptomatic and drug-refractory patients with little or no left atrium enlargement (class IIA, level of evidence C). In the same document, the AHA/ACC/ESC task-force committee suggested a comparison, which is still lacking, of the main effects of AF catheter ablation with a strategy of rate-control, such as A&P for having a comprehensive evaluation of both the therapies.16

Accordingly, we conducted an observational study in order to analyse the effects of A&P and PVI on some clinical endpoints in a large series of consecutive patients with drug-refractory AF. The choice between the two therapeutic strategies was mainly related to clinical characteristics of individual patients, after discussion of the potential benefits and risk percentages. The comparison between the cohorts of patients treated with A&P and AF ablation identified important differences in the baseline clinical profile, probably related to selection criteria. A&P group included older patients with higher prevalence of underlying heart disease, more advanced functional classes, persistent or permanent AF, and lower mean value of LVEF. In the study by Hsieh et al.,15 including only elderly patients with medically refractory paroxysmal AF, similar differences between the groups of patients treated with A&P and AF ablation have been reported for prevalence of cardiovascular disease, not for LVEF, history of heart failure, and NYHA class. The number of ineffective antiarrhythmic drugs used before ablation was instead similar to our study in both the groups. Moreover, only single-chamber VVI or VVI-R pacemakers where utilized in the A&P group, only an average of 1.3 PVs per patient were isolated in the PVI group and anticoagulation therapy was utilized in 3% of both the groups.15

In our experience, total procedure, fluoroscopy, and RF delivery times were significantly longer in the group of patients treated with a complex therapy such as AF ablation. However, the results appear very similar to other large experiences1012 and literature reviews.9,23 The acute success of A&P and AF ablation by PVI was not different in the two groups and similar to other large studies2,3,7,9,11,18 and reviews.6,23 Major periprocedural complications were lower for A&P than for PVI but did not show any particular sequelae. During the follow-up, all the patients treated with A&P were free of symptoms related to AF recurrence, while a good clinical control has been obtained in 79% of patients who underwent PVI. However, persistent use of antiarrhythmic drugs has been reported in 47% of patients after AF therapy, and only in 9% of the cases after A&P. The development of persistent or permanent AF after AVJ ablation was significantly higher in the A&P group. According to our data, the development of permanent AF in the absence of antiarrhythmic drugs was reported in 69% of patients treated by AVJ ablation in the study by Hsieh et al.15

Finally, total and cardiac deaths (16 and 6% after 24 months, respectively) were observed only in the A&P group and appear similar to single centre large experience3 and meta-analysis.6 The baseline older age associated to a worse LV function and higher rate of persistent AF in comparison with patients treated with PVI may have contributed to the different total and cardiac mortality. Indeed, the adjusted, frailty-based analysis performed, down-corrected the significance of the group effect on mortality, which turned out to be not significant after such adjustments.

A few observational27,28 and randomized studies2933 have recently validated the efficacy of AVJ ablation plus biventricular pacing in selected patients with drug-refractory AF, chronic heart failure, and severe LV dysfunction. In the prospective randomized PAVE study30 biventricular pacing provided a significant improvement in the 6-min hallway walk test and EF compared with right ventricular pacing, and the favourable effect of cardiac resynchronization appeared to be greater in patients with reduced LVEF (≤45%) or with symptomatic heart failure (NYHA class II–III). In the OPSITE29,32 study rhythm regularization obtained with A&P improved quality-of-life indexes and exercise capacity with right ventricular, LV, and biventricular pacing. Surprisingly, biventricular pacing was better than right ventricular pacing only in the subgroup of patients with preserved LVEF and absence of native left bundle branch block. A small pilot study compared PVI vs. AVJ ablation plus biventricular pacing among patients of 60 years of age with drug-refractory paroxysmal or permanent AF, and mean 27% value of EF at baseline. At 6 months, 72% of patients who underwent PVI were free from AF, and <30% of patients had second procedure. Successful PVI significantly improved quality of life, EF, as well as 6-min walking test when compared with AVJ ablation with biventricular pacing.33

Study limitations

  1. The allocation to A&P or PVI was not randomized, but related to clinical selection and patient preference. During the study period international guidelines for the management of patients with drug-refractory AF were not available in order to correctly identify the best therapeutic strategy. We addressed the major limits of the possible unbalance in baseline clinical characteristics, with the potential effect of biasing the comparison of the two cohorts, using up-to-date statistical approaches.22 In particular, we compared A&P and PVI with multivariable frailty model, thus adjusting the baseline unbalance adding gender, age, LVEF, NYHA class, organic heart disease, and AF type as confounders in the model. This is a severely conservative modelling strategy, as shown by the evaluation of the mortality among groups, which turned out to be not significant in the multivariable model proposed.

  2. The most effective AF ablation method to eliminate the arrhythmia was not well-defined when we started the observational study. We chose the electrophysiologically-guided PVI method, which was integrated in the last 76 out of 144 patients with electroanatomical CARTO mapping. However, the acute and long-term results obtained in our series are similar to what was reported in multicentre experience and large literature reviews.9,23

  3. In patients with persistent or permanent AF, we limited the RF delivery to the atriopulmonary junction and we did not create linear lesions in special area of the left and right atrium or at sites with fragmented potentials, as suggested recently.8,25,26,34 The application of these new methods could have improved the AF ablation outcome in that subgroup of patients.

  4. Finally, we have considered the long-term outcome of both the groups of patients only by clinical evaluation, ECGs, Holter recordings, and in the A&P group by pacemaker diagnostic function (atrial and ventricular histograms, mode-switch, and atrial tachycardia episodes). Quality-of-life indexes and exercise performance were not utilized to objectively assess patient outcome in this series. However, such assessments are well documented in the literature.3,4,35


In this series patients treated with A&P and PVI for drug-refractory AF showed significant differences in clinical profile and demographics probably related to the selection bias and procedural aspects. Using an up-to-date statistical approach (multivariable frailty model), AVJ ablation plus single- or dual-chamber therapy resulted in a favourable long-term success on AF-related symptoms, but was associated with higher incidence of permanent AF in comparison with AF ablation. Our observational study suggests that A&P might be a reasonable alternative to PVI at least in the selected cases. However, randomized studies comparing during longer follow-up, the ‘rate- vs. rhythm-control’ ablative procedures appear necessary to identify the best strategy.17


The authors acknowledge Dr. Paolo Esente for the manuscript revision and important suggestions.

Conflict of interest: none declared.


View Abstract