Predictive value of the cholesterol-HDL-glucose (CHG) index for recurrence after radiofrequency catheter ablation of atrial fibrillation: a retrospective cohort study

Radiofrequency catheter ablation (RFCA) is a primary rhythm-control therapy for atrial fibrillation (AF), yet high post-procedural recurrence limits its long-term benefit. Cost-effective preoperative biomarkers are required to refine individualized risk stratification. This study evaluated the predictive performance of the cholesterol‑HDL‑glucose (CHG) index for AF recurrence after RFCA. This retrospective cohort included 625 consecutive patients with non-valvular AF who received initial RFCA (2020–2024) and had complete 12-month follow-up. Subjects were divided into recurrence and non-recurrence groups. Multivariate Cox regression identified independent predictors of recurrence. ROC curves quantified the CHG index’s predictive power and optimal cut-off. Kaplan–Meier, restricted cubic spline (RCS), DeLong test, continuous NRI, IDI, and DCA were adopted to assess its linear association and incremental prognostic value. Subgroup interaction analyses further validated predictive consistency across clinical subgroups. The 12-month AF recurrence rate was 28.2% (176/625). The CHG index emerged as an independent recurrence risk factor, with an AUC of 0.698 (95% CI 0.651–0.745) and an optimal cut-off of 5.20. Higher CHG index corresponded to elevated cumulative recurrence risk (Log-rank P < 0.001), and RCS confirmed an approximate linear correlation between CHG index and recurrence risk. Adding the CHG index to the baseline clinical model significantly improved prediction (IDI = 0.092, continuous NRI = 0.542; both P < 0.001). Subgroup interactions existed for age, hypertension, and diabetes (all interaction P < 0.05). The CHG index independently predicts post-RFCA AF recurrence. Adding the CHG index to traditional risk models improves discrimination and risk stratification, supporting its clinical utility for personalized recurrence evaluation.