Validation of the SCALE-CryoAF risk model to predict very late return of atrial fibrillation after cryoballoon ablation

Background: To date, few risk models have been validated to predict recurrent atrial fibrillation (AF) >1 year after ablation. The SCALE-CryoAF score was previously derived to predict very late return of AF (VLRAF) >1 year following cryoballoon ablation (CBA), with strong predictive ability. In this study, we aim to validate the SCALE-CryoAF score for VLRAF after CBA in a novel patient cohort.

Methods: Retrospective analysis of a prospectively maintained single-center database was performed. Inclusion criteria were pulmonary vein isolation using CBA 2017-2020. Exclusion criteria included prior ablation, <1-year follow-up, lack of pre-CBA echocardiogram, additional ablation lesion sets, and documented AF recurrence 90-365 days post-CBA. The area under the curve (AUC) of SCALE-CryoAF was compared to the derivation value and other established risk models.

Results: Among 469 CBA performed, 241 (61% male, 62.8 ±11.7 years old) cases were included in analysis. There were 37 (15.4%) patients who developed VLRAF. Patients with VLRAF had a higher SCALE-CryoAF score (VLRAF 5.4 ± 2.7; no VLRAF 3.1 ± 2.9; p<0.001). SCALE-CryoAF was linearly associated with VLRAF (y=14.35x-11.72, R²=0.99), and a score > 5 had a 32.7% risk of VLRAF. The SCALE-CryoAF risk model predicted VLRAF with an AUC of 0.74, which was similar to the derivation value (AUC_derivation: 0.73) and statistically superior to MB-LATER, CHA2DS2-VASc, and CHADS₂ scores.

Conclusions: The current analysis validates the ability of SCALE-CryoAF to predict VLRAF after CBA in a novel patient cohort. Patients with a high SCALE-CryoAF score should be monitored closely for recurrent AF >1 year following CBA.