Risk model for predicting complications in patients undergoing atrial fibrillation ablation

Santosh K. Padala, Sampath Gunda, Parikshit S. Sharma, Le Kang, Jayanthi N. Koneru, Kenneth A. Ellenbogen

Predictors of complications from AF ablation have been identified in small studies. The combination of risk factors to predict complications after ablation has not yet been explored. To develop a risk score model that predicts complications after AF ablation. The National Inpatient Sample database was utilized to identify 106,105 patients who underwent AF ablation. The study population was split into derivation cohort (DC, 2007-2010, n= 56,658) and validation cohort (VC, 2011-2013, n= 49,447). Multivariate predictors of any complication were identified in DC using regression analysis and a risk score model was developed. The cohorts were divided into 5-groups (risk score in parentheses): group 0 (0), 1 (1-10), 2 (11-20), 3 (21-30) and 4 (31-61). Patients in VC were older, likely to be white, female and had a higher prevalence of co-morbidities. The overall complication rate (6.9% vs. 8.3%, p<0.0001) and in-hospital mortality rate (0.3% vs. 0.5%, p<0.0001) was lower in VC vs. DC. Multivariate analysis yielded 9 predictors for any complication (weightage points in parentheses): CVA (19), CHF (12), coagulopathy (11), renal failure (7), PVD (6), age ≥50 years (2), female (2), COPD (1) and non-white (1). In the overall cohort, risk of complications in groups 0, 1, 2, 3 and 4 was 3.6%, 6.5%, 15.5%, 29.5%, and 45.7% respectively and in-hospital mortality was 0%, 0.2%, 2%, 4.6% and 6.1% respectively. Similar trends were observed in DC and VC. A practical risk score model can be used preoperatively to risk stratify patients undergoing AF ablation.