Background:

Rhythm control therapy for atrial fibrillation (AF) advances more and more to a prognostic rather than symptomatic treatment for AF. Pulmonary vein isolation (PVI) is an effective and safe intervention to maintain sinus rhythm. Thus, there is a growing need to identify patients at risk of experiencing AF recurrence following PVI. Echocardiographic imaging, as well as simple clinical variables such as age, sex and AF subtype are commonly available in most clinical settings.

Purpose:

To assess the risk of recurrence of AF in patients presenting for de novo AF ablation using echocardiographic imaging variables, especially left atrial reservoir strain (LASr) and left atrial volume indexed to BSA (LAVI), as well as clinical variables including age, sex and AF pattern.

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. A baseline transthoracic echocardiography was performed in every patient prior to PVI in sinus rhythm. Patients in AF at presentation, 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. Median follow-up time was 799 day (95% confidence interval 776-831 days). All analysis were censored after a follow-up time of 1 year.

Two uni- and multivariable Cox regression models were calculated. The imaging model included global longitudinal strain of the left ventricle, LAVI, left atrial ejection fraction and LASr. The combined model additionally included sex, age, type of AF, left atrial volume indexed to BSA and the imaging variable with the best p-value.

Results:

132 patients (88 (67%) paroxysmal AF, 79 (60%) male), aged 66 years at median (IQR 55;73) were included. Median CHA₂DS₂-VASc score was 2 (IQR 1;3) and left ventricular ejection fraction was 58% (IQR 52%;64%) in the overall cohort. AF recurrence after 1 year occurred in 30 patients (23%).

Left atrial imaging parameters were assessed in the overall cohort: LAVI was 29.6 ml/m² (IQR 22.9ml/m²;35.7ml/m²), left atrial ejection fraction was 40.6% (IQR 31.9%;49.7%), global longitudinal strain of the left ventricle was -19.6% (IQR -22.1%,-17.8%) and LASr was 27.7% (IQR 18.7%;36.9%). The variable with the best association with AF recurrence after 1 year in the uni- and multivariable Cox-regression including only the imaging variables was LASr (HR 0.96, 95% CI 0.94-0.98, p<0.001). Hazard-ratio per unit increase was 0.96 (95% CI 0.94-0.99, p=0.003) (Figure 1 A+C). LASr was included into the combined model which predicted AF recurrence with a HR of 0.96 (95%CI 0.94-0.98, p<0.001) and 0.97 (95% CI 0.94-0.99, p=0.02) per unit increase (Figure 1 B+D).

Conclusion:

Data of the ASTRA-AF pilot study underline the finding that impaired left atrial reservoir strain is associated with recurrent AF. Independent validation of this finding is warranted.

Figure 1. Cox regression analysis displaying hazard ratios (HR) with 95% confidence-intervals (CI) for the uni- and multivariable analysis in forest plots.