Glycemic Variability and Systemic Immune-Inflammation Index: A Synergistic Predictive Model for Atrial Fibrillation in Type 2 Diabetes

Purpose This study aimed to investigate the association of glycemic variability (GV) and the systemic immune-inflammation index (SII) with Atrial fibrillation (AF) in patients with type 2 diabetes mellitus (T2DM). Methods In this retrospective study of 1,436 T2DM patients, we used multivariable logistic regression and restricted cubic splines (RCS) in the derivation cohort to assess the associations of GV and SII with AF. Robustness was tested via sensitivity and subgroup analyses, while the combined model's discriminative performance was evaluated using receiver operating characteristic curves, interaction analysis, integrated discrimination improvement (IDI), and net reclassification improvement (NRI). The model subsequently underwent external validation in the validation cohort, where its calibration and discrimination were assessed. Finally, decision curve analysis (DCA) was employed to determine its clinical utility. Results In the derivation cohort, multivariable logistic regression confirmed GV (OR:2.13, 95%CI:1.57-2.93; P<0.01) and SII (OR:1.81, 95%CI:1.25-2.62; P<0.01) as independent risk factors for AF, with RCS analyses revealing significant nonlinear relationships. The robustness of these associations was supported by consistent results in subgroup and sensitivity analyses. The combined GV-SII model demonstrated significantly discriminatory power than either marker alone (AUC, GV: 0.709; SII: 0.753; combined: 0.812), reflected by marked improvements in both the NRI and IDI. DCA further affirmed the clinical utility of the combined model. All key findings were successfully replicated in an external validation cohort. Conclusions Both GV and SII demonstrate significant discriminative value for the occurrence of AF in T2DM patients, and their combination yields superior diagnostic performance, enabling the identification of high-risk individuals.