Abstract
BACKGROUND: The clinical factors associated with the recurrence of atrial fibrillation (Af) in patients undergoing catheter ablation (CA) are still ambiguous to date. PURPOSE: 1. To recognize preoperative serologic factors and clinical features associated with Af recurrence after the first ablation treatment. 2. To Develop a Logical Regression Model for Predicting the Likelihood of Recurrence Within 1 Year After the Initial Radio-Frequency Catheter Ablation (RFCA) Therapy. METHODS: Atrial fibrillation patients undergoing RFCA at our institution from January 2016 to June 2021 were included in the analysis (n = 246). A combined dataset of relevant parameters was collected from the participants (clinical characteristics, laboratory results, and time to recurrence) (n = 200). We performed the least absolute shrinkage and selection operator (Lasso) regression with 100 cycles, selecting variables present in all 100 cycles to identify factors associated with the first recurrence of atrial fibrillation. A logistic regression model for predicting whether Af would recur within a year was created using 70% of the data as a training set and the remaining data to validate the accuracy. The predictions were assessed using calibration plots, concordance index (C-index), and decision curve analysis. RESULTS: The left atrial diameter, albumin, type of Af, whether other arrhythmias were combined, and the duration of Af attack time were associated with Af recurrence in this sample. Some clinically meaningful variables were selected and combined with recognized factors associated with recurrence to construct a logistic regression prediction model for 1-year Af recurrence. The receiver operating characteristic (ROC) curve for this model was 0.8695, and the established prediction model had a C-index of 0.83. The performance was superior to the extreme curve in the decision curve analysis. CONCLUSION: Our study demonstrates that several clinical features and serological markers can predict the recurrence of Af in patients undergoing RFCA. This simple model can play a crucial role in guiding physicians in preoperative evaluation and clinical decision-making.