Non-invasive mapping of atrial slow-conduction areas and validation by endocardial isochronal mapping to predict atrial fibrillation recurrence after ablation

Eric Invers-Rubio (Barcelona)1, I. Hernandez (Valencia)2, J. Reventos (Barcelona)1, J.-B. Guichard (Barcelona)1, E. Guasch (Barcelona)1, J. M. Tolosana (Barcelona)1, A. Porta-Sanchez (Barcelona)1, E. Arbelo (Barcelona)1, J. Brugada (Barcelona)1, I. Roca-Luque (Barcelona)1, M. Guillem (Valencia)2, A. Climent (Valencia)2, L. Mont (Barcelona)1, T. Althoff (Barcelona)1

1Hospital Clinic University of Barcelona Arrhythmia Section Barcelona, Spanien; 2Universitat Politècnica de València ITACA Institute Valencia, Spanien


Atrial arrythmogenic substrate is a key determinant of arrhythmia recurrence after pulmonary vein isolation (PVI), and reduced conduction velocities have been linked to adverse outcome. However, a non-invasive method to assess such electrical substrate is not available to date. We have developed and validated a non-invasive mapping system based on electrocardiographic imaging (ECGi) to assess local conduction velocities and their predictive value regarding 12-months arrhythmia-free survival following PVI. 
A novel ECGi-based non-invasive mapping system obviating the need of additional CT- or CMR-imaging was developed. 52 consecutive patients scheduled for AF ablation (PVI-only) and 19 healthy controls were prospectively included and non-invasive mapping to determine local left and right atrial conduction velocities in sinus rhythm. In 36 of the patients high density endocardial mapping of the left atrium was performed using multipolar small-electrode catheters to define areas of slow conduction based on isochronal late activation maps. AF ablation patients were systematically followed with in-person visits and holter ECGs at 2 weeks and 3, 6 and 12 months post ablation.
Mean ECGi-determined atrial conduction velocities were significantly lower in AF patients (54% paroxysmal, 46% persistent AF) than in healthy controls (1.45±0.15 versus 1.64±0.15 m/s; p<0.0001). Differences were even more pronounced in a regional analysis based on a bi-atrial model with 15 pre-defined segments (Fig. A). In this analysis we focused only on the segment with the slowest average conduction velocity in a given patient, i.e. the region that was most likely to harbour arrhythmogenic substrate.
Multivariable regression analyses and c-statistics found this average conduction velocity of the „slowest“ segment (AUC 0.74) to predict 12-months arrhythmia recurrence after PVI better than previously proposed predictors including left atrial size (AUC 0.71) or the left atrial extent of late gadolinium enhancement (MRI) (AUC 0.70). 
A ROC analysis found a conduction velocity threshold of 0.78 m/s to best discriminate PVI responders from non-responders. Patients without slow-conduction areas based on this definition (mean conduction velocity <0.78 m/s) showed significantly higher 12-months arrhythmia-free survival than those with one or more slow-conduction areas (88.9% versus 48.0%, p=0.002).
For validation we compared slow-conduction areas determined by non-invasive mapping (<0.78 m/s) with slow-conduction areas as defined by invasive isochronal mapping. Interrater reliability testing revealed good agreement between non-invasive mapping-determined slow-conduction areas and those defined by isochronal mapping (kappa 0.63, Fig. C).
ECGi-based non-invasive mapping reliably identifies areas of slow conduction. The presence of ECGi-determined slow-conduction areas well discriminates PVI-responders from non-responders. Such non-invasive assessment of electrical arrhythmogenic substrate may guide treatment strategies and could be a step towards personalised AF therapy.
Diese Seite teilen