Abstract
PURPOSE: Patients with chronic obstructive pulmonary disease (COPD) and atrial fibrillation (AF) face elevated ischemic stroke (IS) risk. This study assessed the triglyceride-glucose (TyG) index as a metabolic marker for IS risk in this population. PATIENTS AND METHODS: This retrospective analysis included 710 hospitalized patients with COPD and AF (2014-2024). The TyG index (ln[fasting triglycerides (mg/dL)×fasting glucose (mg/dL)/2]) measured at admission was the exposure; new-onset IS occurring during the index hospitalization (median duration: 9 days, IQR: 6-13) was the outcome. Univariate and multivariate logistic regression (adjusting for age, sex, smoking, blood pressure, diabetes, lipids, prior stroke) identified associations. Nonlinearity was assessed using generalized additive models (GAM). Predictive performance was evaluated via ROC analysis (AUC). RESULTS: IS occurred in 32 patients (4.5%). Unexpectedly, and in contrast to findings in the general population, higher TyG index was associated with lower IS risk in univariate analysis (OR=0.49, 95% CI:0.26-0.94). After full adjustment, each unit increase in TyG index was associated with lower IS risk (aOR=0.24, 95% CI:0.10-0.57, P=0.001). However, the wide confidence interval and limited events (n=32) indicate this large effect estimate should be interpreted cautiously, as it may reflect statistical instability. The finding suggests a potential metabolic paradox in this population. GAM confirmed a linear inverse relationship after adjusting for stroke history (OR=0.50, 95% CI:0.30-0.90, P=0.032). The TyG index predicted IS with an AUC of 0.614 (95% CI:0.513-0.715). CONCLUSION: Contrary to observations in the general population, in hospitalized COPD and AF patients, a higher TyG index was associated with lower short-term ischemic stroke risk. However, this inverse association could reflect either a true inverse metabolic relationship or reverse causality whereby low TyG marks disease severity (malnutrition, poor metabolic reserve, chronic inflammation). Lack of comprehensive disease severity and nutritional assessments limits causal inference. The TyG index may serve as a potential biomarker in hospitalized patients, but population-based studies are essential to address selection bias.