The effect of non-insulin-based insulin resistance indices on the prediction of recurrence in patients with atrial fibrillation undergoing radiofrequency catheter ablation

Abstract Background Atrial fibrillation (AF) is acknowledged as a disease continuum. Despite catheter ablation being recommended as a primary therapy for AF, the high recurrence rates have tempered the initial enthusiasm. Insulin resistance (IR) has been established as an independent predictor for the onset of AF. However, the correlation between non-insulin-based IR indices and late AF recurrence in patients undergoing radiofrequency catheter ablation remains unknown. Methods A retrospective cohort of 910 AF patients who underwent radiofrequency catheter ablation was included in the analysis. The primary endpoint was late AF recurrence during the follow-up period after a defined blank period. The relationship between non-insulin-based IR indices and the primary endpoint was assessed using multivariate Cox hazards regression models and restricted cubic splines (RCS). Additionally, the net reclassification improvement and integrated discrimination improvement index were calculated to further evaluate the additional predictive value of the four IR indices beyond established risk factors for the primary outcome. Results During a median follow-up period of 12.00 months, 189 patients (20.77%) experienced late AF recurrence, which was more prevalent among patients with higher levels of IR. The multivariate Cox hazards regression analysis revealed a significant association between these IR indices and late AF recurrence. Among the four indices, METS-IR provided the most significant incremental effect on the basic model for predicting late AF recurrence. Multivariable-adjusted RCS curves illustrated a nonlinear correlation between METS-IR and late AF recurrence. In subgroup analysis, METS-IR exhibited a significant correlation with late AF recurrence in patients with diabetes mellitus (HR: 1.697, 95% CI 1.397 − 2.063, P