Over the past decade, catheter ablation of atrial fibrillation (AF) has become a routine procedure in clinical practice. Operator learning curves have been paralleled by a considerable improvement in technology and systematization of the technique.1, 2, 3 There is, however, little published data on how modifications to the technique, the acquisition of experience by operators, or the evolution of technology affects outcomes.

Due to the rapid expansion of indications for ablation and the growing demand for AF ablation,4, 5, 6, 7, 8 there is a need to identify and establish safe procedures, and to analyze the actual rate of complications occurring in different centers.

The aim of this study was to analyze outcomes and complications in patients who underwent percutaneous ablation of pulmonary veins (PV) in our center over the past 7 years. We also aimed to identify possible predictors of success in the treatment of AF and potential sources of complications.

Between October 2002 and December 2009, a total of 726 percutaneous PV ablation procedures were performed in 542 patients using 3-dimensional non-fluoroscopic mapping (CARTO® or NAVX®).

All patients underwent transesophageal echocardiography in the 48h prior to ablation to rule out the presence of thrombi. In 74.2% of the procedures performed, patients also underwent computed tomography (CT) or magnetic resonance imaging (MRI) of the left atrium (LA) and PVs. Images obtained were integrated with those from the electroanatomic mapping system to achieve better spatial resolution and anatomic definition during ablation.

Oral anticoagulation was discontinued 3 days before ablation, and administration of low molecular weight heparin was initiated the day before the procedure.

Beginning in January 2008, all procedures were performed using a protocol for anticoagulation and conscious sedation (Table 1). Patients following this anticoagulation and sedation protocol constituted group B, while patients in the group that underwent PV ablation before 2008 constituted group A.

Table 1. Anticoagulation and Conscious Sedation Protocol Applied Systematically in the Ablation of Atrial Fibrillation Since January 2008.

ACT, activated coagulation time; INR, international normalized ration; LMWH, low molecular weight heparin; NSS, normal saline solution; UFH, unfractionated heparin.

All patients signed informed consent before the procedure and the study was approved by the center's ethics committee.

Ablation was performed percutaneously via the femoral vein with monitoring of arterial oxygen saturation and invasive monitoring of blood pressure. After a double transseptal puncture, the LA was accessed and the ablation catheter positioned there. A circular catheter (Lasso®, Biosense-Webster Lasso, or Inquiry Optima®, St. Jude Medical) has been used since 2006 for recording and stimulation. Unfractionated heparin sodium (UFH) was then administered, based on established strategy. Three-dimensional mapping of the LA and adjacent structures was carried out using CARTO® (Biosense Webster) or NavX® (St. Jude Medical) and, whenever possible, MRI or CT scan was also used to optimize anatomical reconstruction. Radiofrequency (RF) was applied with an 8mm ablation catheter or a 3.5mm irrigated tip catheter to a target temperature of 55°C or 45°C and a maximum output of 60 or 40W, respectively. Continuous RF lesions were used to encircle the ipsilateral PVs. Ablation lines were also created on the posterior wall, the LA roof, and the mitral isthmus. We ablated areas with high fragmentation of local electrograms in some subgroups of patients, based on AF type or atrium size. The fact that our center participated in randomized trials during the study period produced changes to the methodology for creating lines or ablating fragmented electrograms in certain patient groups.

The procedure aimed to reduce local PV electrogram voltage to <0 15mv in patients which a circular catheter was not used or to eliminate pv potentials when catheters were further aim verify the bidirectional block between la and blocking of roof line confirmed by presence double caudocraneal activation posterior wall isolation disappearance potential lack atrial capture with local there no systematic check mitral isthmus without history left flutter

In group A, midazolam and fentanyl were administered during the procedure based on the operator's judgement, while in group B a systematic sedation approach was taken, using dolantina, midazolam, and fentanyl (Table 2).

Table 2. Baseline Patient Characteristics and Differences Between Subgroups According to Whether Ablation Was Completed Before or After January 2008.

BMI, body mass index; EF, ejection fraction; LA, left atrium; LVDD, left ventricular diastolic diameter; LVSD, Left ventricular systolic diameter; OSAS, obstructive sleep apnea syndrome.

a Patients with an apnea-hypopnea index >10 on the Berlin questionnaire or patients who, at baseline, were using positive-pressure noninvasive ventilation apparatus, or continuous positive airway pressure at night 5 .
b High performance athlete, patients regularly practicing vigorous prolonged sporting activity, at least 3 h/weekly for >2 years 7 .

The anticoagulation strategy in group A was limited to initial bolus administration of UFH with variable monitoring of activated clotting time (ACT) and administration of variable amounts of heparin based on operator judgement. Anticoagulations of <250ms were accepted in group b an initial bolus of 5000 or 6000 iu ufh was given depending on the patient s weight followed by regular and established monitoring act administration dose heparin until values between 250ms 300ms reached

After ablation, and following the guidelines of the Heart Rhythm Society/European Heart Rhythm Association/European Cardiac Arrhtymia Society expert group,4, 5 anticoagulation was maintained in all patients for at least 3 months, and continued or discontinued based on the CHADS2 risk score. The decision to continue or discontinue antiarrhythmic treatment during follow-up was made by the cardiologist. Patient follow-up included visits every 3 to 6 months for at least 1 year. At each visit a surface ECG and 48-hour Holter-ECG was carried out. Between 6 and 9 months, a transthoracic echocardiogram and a CT scan or MRI of the PVs was performed to rule out late complications.

Arrhythmia recurrence was defined as any episode of atrial tachyarrhythmia lasting >30s recorded beyond the first 3 months after ablation. The first 3 months after ablation were considered the wash-out or window period and any arrhythmic events occurring during that time were not counted as recurrences.4

Clinically relevant complications associated with the ablation procedure were recorded. Major complications were defined as those which were life-threatening, caused permanent damage, or required therapeutic intervention and a prolonged stay in hospital.9

Continuous variables are expressed as means ± standard deviations and categorical variables as percentages. Differences in the baseline characteristics of patients who received ablation before and after implementation of the anticoagulation and sedation protocol were analyzed using the Student t test and χ2 statistics. The same methods were used to test for differences between patients in which the procedure was successful and those in which it was not, and in patients with and without clinically relevant complications. To determine the success of the technique, we assessed arrhythmia-free survival using the Kaplan-Meier method and a Cox regression model for the multivariate analysis of all significant factors. Finally, to determine which factors independently predicted complications, all those which were significant in univariate analysis were included in a binary logistic regression model and their odds ratio (OR) estimated. Statistical significance was set at P<.05 in all analyses.

From October 2002 until December 2009, a total of 726 PV ablation procedures was performed in 542 patients using nonfluoroscopic mapping. The baseline characteristics of the population are shown in Table 2. Overall, 77% of patients were male, mean age was 53 years, 21.2% had structural heart disease, and 42.2% had hypertension. Mean PV diameter in the LA was 42mm and left ventricle ejection fraction was preserved in most patients. Paroxysmal AF was present in 51.3% of patients, 31.3% had persistent AF, 13.4% had long-standing persistent AF, and 4% had left atrial flutter. At least a second procedure was required in 153 patients (28.2%), giving a total of 184 reablations.

After 24 months of follow-up, the overall probability of success after only 1 ablation procedure was 41.1% (51.6% of patients in group B compared to 35.2% in group A, P=.001). That percentage rose to 60.9% after a repeat procedure. Among patients with a recurrence, 75.3% had predominantly AF and 24.7% had atypical atrial flutter. Table 3 shows the characteristics of patients with and without recurrence of arrhythmia after the first ablation. As shown by the survival curves after the first procedure, arrhythmia recurred more frequently in patients with paroxysmal AF (Figure 1A), hypertension (Figure 1B), dilated LA (Figure 1C), and obstructive sleep apnea syndrome (OSAS) (Figure 1D). Table 4 shows the predictors of recurrence: hypertension and OSAS were found to independently predict recurrence after a first ablation procedure. Patients with a dilated LA (>44mm) and paroxysmal AF showed a tendency toward recurrence but the trend was not statistically significant in multivariate analysis. The presence of structural heart disease or a mitral line did not independently predict recurrence.

Table 3. Differences in Baseline Characteristics Between Patients With and Without Recurrence After a First Procedure.

BMI, body mass index; EF, ejection fraction; EPS, electrophysiological study; LA, left atrium; PAF, paroxystic atrial fibrillation; ML, mitral line; OSAS, obstructive sleep apnea syndrome; RF, radiofrequency; RL, roof line.
Group A, prior to January 2008; Group B, after January 2008.

Figure 1. A: survival analysis (Kaplan-Meier) arrhythmia-free follow-up after an initial ablation procedure, according to the type of atrial fibrillation. B: survival analysis (Kaplan-Meier) arrhythmia-free follow-up after an initial ablation procedure in hypertensive and normotensive patients. C: survival analysis (Kaplan-Meier) arrhythmia-free follow-up after an initial ablation procedure in patients with left atrial diameter >44 or <45mm d: survival analysis kaplan-meier arrhythmia-free follow-up after an initial ablation procedure in patients with and without obstructive sleep apnea syndrome af atrial fibrillation ht hypertension la left atrium osas

Table 4. Multivariate Analysis: Independent Predictors of Arrhythmia Recurrence.

95%CI, 95% confidence interval; AF, atrial fibrillation; HR, hazard ratio; LA, left atrium; OSAS, obstructive sleep apnea syndrome.

In total, 61 clinically relevant complications (8.4%) were recorded, 48 in group A (11.5%) and 13 in group B (4.2%). The difference between the groups was statistically significant (P=.002) (Table 5). Overall, 31 complications (4.2%) were classified as serious; this type of complication was also significantly more common in group A (26, 6.2%) than in group B (5, 1.6%; P=.002). Thirty complications (4.1%) were classified as minor and there were no statistically significant differences between groups for this type of complication (22 in group A, 5.2%, and 8 in group B, 2.6%; P=.129). The observed differences between subgroups were mainly due to a reduction in embolic complications (coronary and cerebrovascular) in Group B, as shown in Table 5. We recorded a total of 9 coronary air embolisms (1.2%), all of which were transient; in all cases, complete resolution of clinical angina and ST segment elevation in the ECG was achieved with intravenous nitroglycerin. Of these 9 recorded cases of coronary air embolism, 8 (1.9%) occurred prior to the implementation of the protocol (group A) and 1 (0.3%) afterwards (group B). We also recorded 9 cerebral transient ischemic attacks (TIAs). Other major complications were infrequent, with the exception of cardiac tamponade (7, 1%). There were 4 (0.6%) femoral pseudoaneurysms requiring surgical repair of the artery. Of particular note was a rupture of the mitral subvalvular apparatus due to entrapment of the circular catheter during electroanatomical mapping of the LA; in this case, urgent surgical repair of the mitral valve was required. Of further note was a cardiac perforation due to tearing of the venoatrial union of the left superior PV during mapping; this required urgent pericardiocentesis and surgical suture of the tear. With regard to minor complications, there were no differences in the need to interrupt the procedure due to transseptal puncture or aortic pericardial puncture (without cap) (3.8% in group A and 2.3% in group B; P=.29). We recorded 3 cases of pericarditis with nonsevere associated pericardial effusion (0.4%), 2 (0.3%) cases of esophagitis (transient dysphagia with retrosternal burning pain without evidence of atrial-esophageal fistula), and 2 patients (0.3%) had significant stenosis of the PV (defined as a reduction to less than 50% of the vascular diameter of at least one PV) without clinical consequences. The only factors showing a statistically significant association with higher prevalence of any type of complication were protocolization of the procedure, the use of a circular catheter (both protective), and being female (Table 6). Protocolization of the procedure (OR=0.406, 95% confidence interval [95%CI], 0.214 to 0.769; P<.006) and being male (OR=0.503, 95%CI, 0.275 to 0.919; P<.026) were independent predictors of the absence of complications. Although procedures in group B took slightly longer (162±48 min in group B compared to 131±45 min in group A, P<.001) and RF application times were longer (3221±984ms in group B compared to 2307±1062ms in group A, P<.001), these factors were not predictors of complications.

Table 5. Differences in Complications Associated With Pulmonary Vein Ablation According to Whether the Procedure Was Performed According to a Protocol or Not (Group A: Procedures Performed Before January 2008, Without a Protocol; Group B: Procedures Performed After January 2008, Using a Protocol).

PE, pulmonary embolism; PV, pulmonary veins; TIA, transient ischemic accident.
Data are expressed in no. (%).

Table 6. Differences in Patients and Procedure Characteristics of Cases With Complications. Univariate Analysis.

AF, atrial fibrillation; LA, left atrium; OSAS, obstructive sleep apnea syndrome.

The increasing prevalence of AF and growing demand for AF ablation means that increasing numbers of centers are performing the treatment. AF ablation is a complex procedure and its complications, although infrequent, can be serious. The outcomes presented in several series vary widely and depend critically on the center's experience and the type of AF. The high published rates of success in treating paroxysmal AF (over 80% in the series first published by pioneering groups) have decreased to under 70% in more recent series.10, 11, 12 The results also vary widely if we compare success rates for the same group over time, from 6 month success rates of <60 to 1 year recurrence-free rates of 90%. 10 , 13 Such differences can be attributed to the learning curve and accompanying technological advancement. The same occurs in persistent AF where success rates range from 50% to 75%, depending on the center, technique, and methodology for the detection of recurrences. 12 , 14 In our series, the overall success rate for ablation was 60.9%, which is somewhat lower than those published by some of these groups. This higher percentage of recurrences in our group may have been due to the use of a very strict definition of recurrence. We considered that any registered arrhythmic episode of over 30s duration from the third month of follow-up onwards should be recorded as a recurrence, regardless of whether it was symptomatic or not or whether the patient was taking anti-arrhythmic drugs. We also performed 48-h (occasionally 7-day) Holter monitoring at 3, 6, and 12 months and insistently recommended that patients obtain ECG recordings if they suffered any symptoms suggestive of arrhythmia. In our center, all procedures included in the series were performed using nonfluoroscopic mapping of the LA together with 3-dimensional CT or MRI images obtained prior to the procedure. The latter were especially useful in dealing with complex anatomies. We also used an irrigated ablation catheter in over 80% of cases. The main objective was to electrically isolate the PVs by applying RF at the venoatrial junction. Previously published studies have shown that PV ablation is less effective in hypertensive patients with dilated LA, nonparoxysmal AF, and OSAS 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22; in our series, we found that hypertension and OSAS were the most powerful independent predictors of recurrence. We also observed that patients with linear mitral isthmus ablation had a higher number of recurrences. Incomplete realization, without verification of the blockade, could have led to a proarrhythmic outcome. For that reason, over the last 2 years the vast majority of procedures have been performed without the mitral line. On the other hand, we found that using a circular catheter to map the LA was associated with a lower rate of total complications. Nevertheless, as this approach was infrequently used in group A and systematically introduced in group B there could be a time effect; whether or not it actually reduces complications is therefore not so clear-cut.

Although operator learning curves and technological advances have led to lower complication rates, we still had a 4.2% rate of major complications in our series. That is similar to the rate observed in the last worldwide survey on the methods, efficacy, and safety of catheter ablation for human AF.23

The majority of complications occur during or immediately after the procedure. Cardiac tamponade is still the most common, potentially life-threatening complication and in most cases is related to the transseptal puncture procedure. In our series, this complication did not decrease significantly over time, although this was probably due to the incorporation of a new operator in the last 2 years of the series; learning curves are highly important with regard to this complication. In our study, of the 61 complications observed, 79% derived from the transseptal puncture (with or without tamponade) and cerebral or coronary embolic events. However, after the introduction of the protocol for periprocedural anticoagulation and conscious sedation, we observed a reduction in thromboembolic complications, from 8 before the introduction of the protocol to 1 after. Before introduction of the protocol, the procedure was performed by applying an initial UFH bolus, variably monitoring the ACT, and administering variable doses of UFH based on operator judgement, all of which entailed accepting suboptimal levels of anticoagulation ( <250s under the new protocol goal is to maintain act levels of 250ms throughout the procedure.

Of course, careful handling of the pods and continuous irrigation systems is particularly important to avoid thrombus formation and introduction of air into the system. In large part, these are responsible for many of the cardioembolic events that occur, particularly coronary events. In this sense, training nurses to monitor anticoagulation levels and irrigation systems, as well as to detect complications early, is essential.

The significant reduction in complications over the years is clearly due to various factors. On the one hand, systematization of anticoagulation during the procedure has helped to achieve optimal and stable levels of anticoagulation, which could explain the lower rate of embolic complications from 2008 onwards. Likewise, ongoing analysis of complications and the exhaustive search for potential causes has contributed to the education of both medical and nursing staff. The learning curve has led to considerable attention being paid to catheter manipulation and the identification of potential sources of complications which no doubt further helps to prevent their occurrence.

In 2008, in parallel with the anticoagulation protocol, we also introduced a conscious sedation protocol for all patients. This was developed using recommendations from colleagues in the anesthesiology service and may have also helped reduce complications by increasing patient comfort, thereby improving catheter stability and the safety of the procedure.

As this was a prospective, nonrandomized study, it was not possible to determine how much of the improvement in outcomes and complications derived from the 3 factors involved: learning, technological improvements, and the implementation of protocols. On the other hand, as the correlation between patient-perceived symptoms and episodes of persistent arrhythmia is often poor, the way arrhythmic recurrences were monitored and recorded during follow-up was key to evaluating the success of the procedure. In this sense, it is important to point out that outcomes were monitored similarly in both study groups. Finally, the lack of standardization and data relating to technical aspects of the procedures, for example in regard to the use of mitral lines, roof lines, or circular catheters, especially in the earlier stages of the study, could also be considered a limitation. Randomized studies are needed to analyze the predictive value of such factors.

Outcomes and complication rates in PV ablation have improved significantly over the years.

The systematic application of anticoagulation and conscious sedation protocols was associated with improved outcomes and reduced complications in AF ablation. Other factors not measured in this study, such as the operator learning curve and ongoing improvements in technology, may have contributed to these changes.

This article was partially financed by the Carlos III National Health Institute (Instituto Nacional de Salud Carlos III): REDINSCOR RD06/0003/0008.

None declared.

Received 18 March 2011
Accepted 3 August 2011

Corresponding author: Sección de Arritmias, Servicio de Cardiología, Institut del Tòrax, Hospital Clínic Universitari de Barcelona, Villarroel 170, 08036 Barcelona, Spain. lmont@clinic.ub.es