Background:

Recurrence of atrial fibrillation (AF) following rhythm control therapy is common. It is observed in about 70% of cases after cardioversion and between 20% and 45% after catheter ablation (PVI). Early rhythm control was reported to reduce cardiovascular events in AF patients and results of catheter ablation surpass those of medical therapy only. Although many factors are reported regarding the risk of developing AF, factors associated with AF recurrence following PVI are not well known.

Purpose:

To validate an imaging-based score for prediction of AF recurrence in patients within one year after de novo AF ablation for paroxysmal or persistent AF. The score includes patient age, left atrial reservoir strain (LAS_r) and left atrial volume indexed to body surface area (LAVI).

Methods:

The ASTRA-AF pilot study prospectively enrolled n=132 consecutive patients with paroxysmal or persistent AF presenting for de novo PVI between December 2017 and January 2019. Baseline transthoracic echocardiography was performed in every patient prior to PVI in sinus rhythm. Imaging variables and the LASr were analyzed with post processing software (TOMTEC-ARENA, Germany). Patients in AF at presentation, or with poor 2D imaging quality or impaired ejection fraction <50% were excluded. Recurrence of AF was assessed after a time-period of 1 year via 24h Holter-ECG. Analysis was censored after a median follow-up time of 1 year.

Receiver operating characteristics curve (ROC) analysis of the variables age and LAS_r was performed with previously validated cutoffs of 71 years for age and 31.4% for LAS_r. Further, LAVI with the ESC-guideline cutoff value of 34ml/m², indicating a dilated left atrium, was included. The score increased by 1 point for each age >71 years, LAS_r ≤ 31.4% and LAVI > 34 ml/m².

Probabilities per score for recurrence of AF are shown in Kaplan-Meier curves (Figure 1). The Log-Rank test in the plot describes if there is any difference between the curves. Pairwise comparisons between groups were added.

Results:

ASTRA-AF included 132 patients (79 (60%) male, 88 (67%) paroxysmal AF). The ASTRA score could be calculated in 129 (98%) patients. The median age of the cohort was 65 (IQR 55;73) years and 59.7% were male (n=77). Median CHA₂DS₂-VASc was 2 (IQR 1;3) in the overall cohort and median follow-up time was 799 (IQR 776;831) days. AF recurrence after 1 year occurred in 30 (23%) patients. Median left ventricular ejection fraction was 9% (IQR 53%;64%) and median LAVI was 29,5ml/m² (IQR 22.8ml/m²;35.7ml/m²).

Median ASTRA score was 1 point (IQR 0;2) in patients with paroxysmal AF and 1 (IQR 1;2) in patients with persistent AF (p=0.09). Log-Rank test was significant for a difference between the Kaplan Meier curves (Figure 1) and a higher ASTRA score was associated with a higher probability for AF recurrence. Patients with an ASTRA score of 0 points had a low probability (8%) of recurrence, while patients with 3 points had a probability of 58%.

Conclusion:

In the ASTRA-AF pilot study assessing the risk of AF recurrence during 1 year follow-up a risk score (ASTRA-Score) including age, left atrial reservoir strain and left atrial volume indexed to BSA was feasible to assess the probability of AF recurrence following catheter ablation. A greater ASTRA score was related to higher probability of recurrence of AF.

Figure 1. Results of the Kaplan Meier analysis regarding recurrence of AF.