A prediction model based on functional mitral regurgitation for the recurrence of paroxysmal atrial fibrillation (PAF) after post‐circular pulmonary vein radiofrequency ablation (CPVA)

AIM: To construct a prediction model based on functional mitral regurgitation (FMR) in patients with paroxysmal atrial fibrillation (PAF) to predict atrial fibrillation recurrence after the post-circular pulmonary vein radiofrequency ablation (CPVA). METHODS: We retrospectively analyzed 289 patients with PAF who underwent CPVA for the first time. The patients were randomly divided into modeling group and verification group at the ratio of 75:25. In the modeling group, the multivariate logistic regression was used to analyze and construct a prediction model for post-CPVA recurrence in PAF patients, which was then validated in the verification group. RESULTS: (1) After 3-6 months of follow-up, the patients were divided into sinus rhythm group (252 cases) and recurrence group (24 cases); (2) In the modeling group, the age, left atrial diameter (LAD), and the degree of MR (mild, moderate, severe) were higher in recurrence group than that of the sinus rhythm group, and the left atrial appendage emptying velocity (LAAV) was lower in recurrence group (all p < .05). (3) A model for predicting the recurrence of PAF after radiofrequency ablation was constructed in the modeling group. The equation was: Logit(P) = -3.253 + .092 × age + 1.263 × mild MR + 2.325 × moderate MR + 5.111 × severe MR -.113 × LAAV. The area under the curve (AUC) of the model was .889 in modeling group and .866 in verification group, and the difference was not statistically significant (p > .05). CONCLUSION: The prediction model of atrial fibrillation (AF) recurrence after CPVA in PAF patients has good predictive efficacy, specificity, and accuracy.