Cardiovascular events and death after catheter ablation in very old patients with nonvalvular atrial fibrillation

Baseline characteristics

Of the 1,687 patients aged ≥80 years diagnosed with AF, 56 with valvular AF, 6 with end-stage renal failure, 184 who underwent AF ablation before 80 years old, 184 not using DOACs, and 475 using warfarin were excluded. Among the 782 patients with Nonvalvular AF (NVAF) using DOACs, 212 underwent catheter ablation of AF (Ablation group), and 570 received medical therapy only (Medication group) (Figure 1).

In the original total cohort, the median follow-up period was 35 and 36 months in the Ablation and Medication groups, respectively. The mean age, prevalence of hypertension and heart failure, proportion of patients with a history of thromboembolisms, and a CHADS₂ score were significantly lower in the Ablation group than in the Medication group. The creatinine clearance values were significantly higher in the Ablation group than in the Medication group. In the Medication group, a rhythm control strategy using antiarrhythmic drugs was adopted in 57 (10%) patients. Regarding the transthoracic echocardiogram (TTE) parameters, the patients in the Ablation group had a significantly lower left atrium (LA) diameter and higher left ventricular ejection fraction (LVEF) than those in the Medication group (Table 1).

After propensity score matching, 208 patients were extracted from each group (Figure 1). Except for the LA diameter and LVEF, the baseline characteristics were well matched between the 2 groups (Table 2).

AF ablation

Among the 212 patients in the Ablation group, all first-time ablation procedures were successful. During the perioperative period, 9 (4.2%) complications occurred, in which 3 (1.4%) were cardiac tamponades, but all were managed either with treatment or observation. No fatal procedure-related complications occurred (Supplementary Table 1). After the initial ablation, 54 (24%) patients had an AF recurrence. Among them, 17 patients underwent a second ablation session. No complications occurred during the operative period. No patients underwent ≥3 ablation sessions. The incidence of an AF recurrence was similar regardless of the AF type (paroxysmal AF: n = 34/134 (25%) vs. non-paroxysmal AF: n = 20/78 (26%)). The patients with paroxysmal AF underwent a second ablation session more frequently than those with non-paroxysmal AF (n = 13/118 (9.7%) vs. n = 4/78 (5.1%)). The details of the AF characteristics and management including of an AF recurrence, specific procedures in the second ablation session, and anti-arrhythmic drug use in the Ablation group patients for each AF type are shown in Supplementary Table 2.

Outcomes

As the primary outcome, the 3-year incidence of cardiovascular events was significantly lower in the Ablation group than in the Medication group: 24 (13.2%) vs. 43 (23.3%), log-rank P = 0.009 and hazard ratio (HR) 0.52 (95% confidence interval (CI) 0.32–0.86) (Figure 2). Among the cardiovascular events, only the incidence of acute heart failure (AHF) significantly differed (P = 0.021). The incidence of stroke and systemic embolism (SSE) was higher in the Ablation group than in the Medication group, however, there was no statistically difference (Table 3). Among the SSE, the incidence of hemorrhagic stroke was higher in the Ablation group than in the Medication group (n = 5 (2.4%) vs. n = 1 (0.5%)), and, the incidence of ischemic stroke was similar between the groups (n = 4 (1.9) vs. n = 5 (2.4%)).

The percentage of patients with ≥2 cardiovascular events was lower in the Ablation group than in the Medication group (n = 3 (1.4%) vs. n = 8 (3.8%)).

As the secondary outcome, the 3-year incidence of cardiovascular death was also significantly lower in the Ablation group than in the Medication group: 5 (3.0%) vs. 15 (7.8%), log-rank P = 0.019, HR 0.32 (95% CI 0.16–0.88) (Figure 3). No AHF/acute coronary syndrome (ACS)-related deaths occurred in the Ablation group. The incidence of a SSE-related death and sudden cardiac death (SCD) were similar between the groups (Table 3).

The percentage of patients with non-paroxysmal AF at baseline was similar between the Ablation and Medication groups (n = 78 (37.5%) vs. n = 77 (37%)). Although all patients in the Ablation group regained sinus rhythm after ablation, sinus rhythm was not maintained in 5 patients, and AF became a permanent form during the follow-up period. In contrast, 28 paroxysmal AF patients in the Medication group progressed to permanent AF through a natural course. As a result, the percentage of patients with permanent AF during the follow-up was significantly lower in the Ablation group than in the Medication group (n = 5 (2.4%) vs. n = 105 (50.4%), P < 0.001). Among the patients with paroxysmal AF, the incidence of the progression from paroxysmal to permanent AF was significantly lower in the Ablation group than in the Medication group: 2 (2.3%) vs. 28 (24.8%), log-rank P < 0.001, HR 0.03 (95% CI 0.004–0.24) (Figure 4). The number of patients that required pacemaker implantations during the follow-up period was similar between the Ablation group and the Medication group (n = 12 (5.8%) vs. 13 (6.3%)).

Sensitivity analyses of the primary and secondary outcomes in the original total cohort exhibited comparable results (Supplementary Figures 1 and 2). The incidence of a cardiovascular event and cardiovascular death were significantly and numerically lower, respectively, in the Ablation group than in the Medication group (Supplementary Table 3). After adjusting for the risk factors of cardiovascular events, the multivariate Cox proportional hazard model determined that catheter ablation was an independent predictor of a non-cardiovascular event occurrence (Supplementary Table 4). In terms of cardiovascular death, the HR of catheter ablation was 0.45, however, it was not statistically significant (Supplementary Table 5). In the original total cohort, the 3-year incidence of non-cardiovascular death was significantly higher in the Medication group than the Ablation group, however, that was similar between the two groups in the propensity score matched cohort (Supplementary Figures 3 and 4). The details of the non-cardiovascular deaths in the total original and propensity-matched cohorts are shown in Supplementary Tables 6 and 7.

Study design and patients

We conducted a prospective cohort study of consecutive patients with AF aged ≥80 years who received medical treatment from March 2011 to August 2021 at Kagawa Prefectural Central Hospital, Kagawa, Japan. The inclusion criteria were as follows: 1) diagnosed with AF; 2) age ≥80 years at the time of enrollment; and 3) DOAC use at the time of enrollment. The exclusion criteria were as follows: 1) presence of valvular AF; 2) end-stage renal failure (creatinine clearance <15 mL/min/1.73 m² or regular dialysis); and 3) underwent AF ablation before 80 years of age. Valvular AF was defined as moderate or severe mitral stenosis or having undergone a mechanical valve replacement.

We basically decided the indication of AF ablation for patients with symptomatic AF after confirming their preserved cognitive function and a self-supporting ability for performing daily life. We also considered those who were expected to have a high procedural efficacy and safety, such as lesser LA enlargement and comorbidities. For patients with non-paroxysmal AF, we basically considered those with an AF persistence duration of less than 12 months and in their early 80’s as indicated. For patients with asymptomatic AF, we proposed ablation therapy as a treatment option especially for those who had a risk of AF-related events, such as a history of heart failure, if they were expected to have a high procedural efficacy and safety. Aside from those, AF ablation was performed based on the decision of the patients, their families, and their attending physicians.

This study conformed to the principles of the Declaration of Helsinki and was conducted after approval from the Clinical Ethics Committee of the Kagawa Prefectural Central Hospital. Written informed consent was substituted by an opt-out method by announcing the handling of the personal data and right to withdraw consent on the website of the study institution.

AF ablation

DOACs were initiated at least a month before the ablation and continued even after the ablation unless DOAC-related adverse events occurred. Transesophageal echocardiography was performed to confirm the absence of any LA thrombi on the day of admission for the ablation.

AF ablation was performed based on the established method [11]. Briefly, electrical pulmonary vein isolation by radiofrequency ablation was performed using an open-irrigation 3.5-mm tip deflectable catheter (Thermocool; Biosense Webster, CA, USA) with contact-force sensing and a 3-dimensional mapping system (CARTO; Biosense, Webster, CA, USA) or cryoablation using a second generation cryoballoon (Arctic Front Advance; Medtronic, MN, USA). We used heparin to maintain the activated clotting time at ≥300 seconds during the procedure. Regarding the radiofrequency ablation, we conducted a traditional ablation strategy delivering radiofrequency energy at a power output of 25–40 W and with a maximum temperature of 45°C. We performed a force-time integral guided ablation until May 2018 and then an ablation index guided ablation from June 2018 to July 2021. The target value of the force-time integral was 80 gram-seconds with a power output delivery of 35 W on the anterior wall, 30 W on the posterior wall, and 25 W in front of the esophagus using a dragging ablation technique. The target value of the ablation index was 350–400 with 35–40 W on the anterior wall, 300–350 with 30–35 W on the posterior wall, and 250 with 30–35 W in front of the esophagus. We set the ablation index values slightly lower considering the safety. Instead, we set the minimum lesion distance closer (≤3 mm). A posterior wall isolation, superior vena cava isolation, and cavotricuspid isthmus ablation were performed as needed. General anesthesia was induced and maintained with a continuous propofol intravenous infusion using a subglottic device to keep the airway open and an artificial respirator to maintain effective ventilation. We tried to avoid multiple ablation sessions by using low-dose antiarrhythmic drugs when AF recurred. We performed a second ablation session depending on their symptom severity: uncontrolled severe symptoms and frequent occurrence while using drugs. During the second ablation session, we checked for pulmonary vein reconnections of the block lines created during the first ablation session and performed a re-isolation if reconnections were identified. A posterior wall isolation, low-voltage area or trigger site ablation, and linear ablation, including a roof line and mitral isthmus ablation, were performed as needed.

After the ablation, 12-lead ECGs were performed every 3 months at clinical visits. Furthermore, 12-lead ECGs or 24-h Holter ECGs were performed as needed, such as at the time of the symptom occurrence. AF recurrence was defined as an atrial tachyarrhythmia, including AF and atrial tachycardia, which was detected on a 12-lead ECG, or an atrial tachyarrhythmia lasting at least 30 seconds that was documented on a 24-h Holter ECG. A permanent form of AF was defined as an atrial tachyarrhythmia occurring all the time, and more than 3 times, with ECGs, which were evaluated at more than one-month intervals.

Data collection and outcomes

We collected data on the patient characteristics, including the medical history, laboratory data, medication use, and electrocardiography and TTE findings. TTE was performed to measure the LA diameter and LVEF. The LVEF was measured using the disk summation method and expressed as the average over five cardiac cycles if AF persisted. We investigated the outcomes from the medical records at Kagawa Prefectural Central Hospital. We further investigated the outcomes in March 2021 by a mail-in questionnaire or telephone call to the patients, families, and their primary care physicians in the patients who completed clinical visits at Kagawa Prefectural Central Hospital.

The primary outcome was an initial cardiovascular event, and the secondary outcome was cardiovascular death. We defined cardiovascular events as AHF requiring hospitalizations, SSEs, ACS, and SCD. We defined cardiovascular death as death caused by AHF, SSEs, ACS, and SCD.

Statistical analysis

We compared the patient background and outcomes between the Ablation and Medication groups. The primary analysis of this study was in the propensity score matched cohort followed by sensitivity analyses in the original total cohort. We used the age, sex, creatinine clearance, prevalence of hypertension, heart failure, and diabetes mellitus, and history of thromboembolisms and coronary artery disease in the logistic regression model to develop the propensity score for the ablation use. We then constructed a 1:1 matched pair for the Ablation and Medication groups with a tolerance of 0.03 based on the propensity score.

Continuous variables were expressed as the mean and standard deviation or median and interquartile range for their distribution. We used the t-test or Wilcoxon rank-sum test for the between-group comparisons. Categorical variables are presented as numbers and percentages, and the intergroup comparisons were conducted using the Chi-square test. The outcomes are presented as the number of initial events. The Kaplan–Meier survival curve and log-rank test were used to compare the incidence rates of the outcomes of the Ablation and Medication groups considering the censoring due to the end of follow-up without outcomes. The effect was expressed as the HR and its 95% CI with the use of a Cox proportional hazard model.

In the propensity score matched cohort, we compared the percentage of the patients with non-paroxysmal AF at baseline and during the follow-up in each group. Among them, we compared the incidence of the progression from a paroxysmal to permanent form during the follow-up in patients with paroxysmal AF.

We also performed a sensitivity analysis for the primary and secondary outcomes on the original total cohort in the same way as the main analysis. A multivariable Cox proportional hazard model was used to determine the factors associated with the primary and secondary outcomes for the purpose of minimizing the selection bias. The clinically relevant candidate variables for cardiovascular events included catheter ablation, an age ≥85 years, female sex, body weight ≤50 kg, estimated glomerular filtration rate <60 mL/min/1.73 m², heart failure, hypertension, diabetes mellitus, history of a thromboembolism, and non-paroxysmal AF. The candidate variables for cardiovascular death included catheter ablation, and an age ≥85 years, body weight ≤50 kg, heart failure, and non-paroxysmal AF. We also investigated the 3-year incidence of non-cardiovascular death as a sensitivity analysis in the propensity score matched and original total cohorts to confirm the accuracy of the propensity score matching.

Two-sided P-values < 0.05 were considered statistically significant. All statistical analyses were performed using JMP version 15 software (SAS Institute Inc., Cary, NC, USA) and SPSS Statistics version 24 software (IBM Institute Inc. Chicago, IL, USA).