SPISE for predicting the presence of atrial fibrillation in patients with coronary artery disease: a comparative analysis with TyG index

Abstract

Insulin resistance (IR) acts as a common pathogenic mediator underlying both coronary heart disease (CHD) and atrial fibrillation (AF). The Single-Point Insulin Sensitivity Estimator (SPISE index) and triglyceride-glucose (TyG) index are two readily available laboratory surrogates for IR. This study sought to compare their independent predictive capacities for concomitant AF among patients with CHD. Baseline clinical data and laboratory measurements were collected to calculate SPISE index and the TyG index. All enrolled patients were stratified into AF and non-AF groups according to their baseline cardiac rhythm. Intergroup comparisons of the two metabolic biomarkers were conducted. Univariate and multivariate logistic regression models were constructed to identify independent correlates of AF. Restricted cubic spline (RCS) modeling was utilized to explore the nonlinear association between SPISE index and AF. Receiver operating characteristic (ROC) curve analysis was performed to quantify and compare the predictive capacity of SPISE index and the TyG index for AF. Predefined subgroup analyses were further carried out to evaluate potential effect modification across clinical strata. SPISE index levels were significantly higher in the non-AF group relative to the AF group ( P  < 0.001). After comprehensive adjustment for all confounding covariates, SPISE index was independently inversely associated with prevalent AF (OR = 0.627, 95% CI: 0.528–0.746, P  < 0.001). Restricted cubic spline combined with segmented regression analyses identified a significant nonlinear association between SPISE index and AF with a distinct inflection point. ROC curve analysis demonstrated that SPISE index exhibited superior predictive performance for AF risk compared with the TyG index (AUC: 0.646 vs. 0.566). The difference in AUC values between the two markers reached statistical significance according to Delong’s test. Stratified subgroup analyses revealed that the inverse correlation between SPISE index and prevalent AF remained stable across all predefined strata, including sex, age, smoking status, alcohol consumption, hypertension, diabetes mellitus, and BMI. SPISE index represents an independent correlate of prevalent AF in CHD patients. Compared with the TyG index, this convenient insulin sensitivity biomarker possesses superior screening value for identifying CHD individuals with concomitant AF.