Po-01-233 the association between recurrence after atrial fibrillation ablation and biomarkers of cardiac remodeling

Recurrence after AF ablation remains common. Clinical risk scores to predict recurrence, such as the APPLE score, have been developed. However, their ability to accurately predict recurrence is modest and they have not been widely adopted into clinical practice. There has therefore been increasing interest in refining our ability to predict recurrence with biomarkers, though studies have been limited by small sample size and heterogeneity. We aimed to evaluate the association between recurrence after AF ablation and levels of biomarkers associated with cardiac remodeling (full list in Table 1) and to assess whether adding a biomarker to a clinical risk model improves its predictive ability. Participants underwent a de-novo catheter ablation (2011-2021). Serum and plasma were collected prior to ablation and stored. Biomarkers were assayed in collaboration with Roche Diagnostics Laboratory (Penzberg, Germany). Recurrence was defined as ≥30 seconds of an atrial arrhythmia 3-12 months post-ablation. Multivariable logistic regression was performed using biomarker levels along with pre-specified clinical co-variates: APPLE score (which includes age, left atrial diameter, left ventricular ejection fraction, persistent AF, and eGFR), rhythm at the start of the case, sex, height, body mass index, presence of an implanted continuous monitor, year of ablation, and whether additional ablation (e.g., posterior wall) was performed. Models with and without biomarkers were compared using a likelihood ratio test. Internal validation was performed using bootstrapping. A total of 1,873 participants were included in our study. Median age at ablation was 66.1 (IQR 58.9-72.6). There were 1,189 (63.5%) male participants and 1,008 (53.8%) had paroxysmal AF. A multivariable logistic regression (Figure 1) showed an association between recurrence and levels of Ang2 (odds ratio (OR) 1.08 (95% CI 1.02-1.15), p=0.007), IGFBP7 (OR 1.1005 (1.0002-1.009), p=0.04), and IL6 (OR 1.02 (1.003-1.03), p=0.02). The area under the receiver operating curve (AUC) for the logistic model including the biomarker ranged from 0.711-0.714 for the different biomarkers (Table 1). There was no statistically significant difference between the AUC of models with a biomarker compared to without (AUC 0.711). Ang2, IGFBP7, and IL6 were independently associated with recurrence after AF ablation. The addition of biomarkers to a clinical model does not appear to significantly improve prediction of recurrence.Tabled 1Multivariable Logistic RegressionBiomarkerMultivariable logistic regression model for original cohort (N=1873)Odds Ratio (95% CI)P-valueAUC of model with biomarkerBone morphogenetic protein 10 (BMP10)1.12 (0.91-1.4)0.280.711 (0.689-0.736)Angiopoietin 2 (Ang2)1.08 (1.02-1.15)0.0070.714 (0.689-0.739)Fibroblast growth factor 23 (FGF23)1.00 (0.99-1.00)0.090.712 (0.686-0.737)Insulin-like growth factor binding protein 7 (IGFBP7)1.005 (1.0002-1.009)0.040.711 (0.686-0.737)Myosin binding protein C3 (MYBPC3)1.003 (0.999-1.0007)0.130.712 (0.687-0.737)Growth differentiation factor 15 (GDF 15)1.00007 (0.999-1.0002)0.090.712 (0.686-0.737)Interleukin 6 (IL6)1.02 (1.003-1.03)0.020.713 (0.688-0.739)N-terminal pro brain natriuretic peptide (NT proBNP)1.00005 (0.999-1.0001)0.090.712 (0.687-0.737)High sensitivity troponin T (TnT-HS)1.0004 (0.999-1.001)0.420.711 (0.686-0.736) Open table in a new tab