Abstract Background Atrial fibrillation (AF) increases the risk of stroke and thromboembolism; and appropriate risk stratification is central for better targeted patient management. The triglyceride-glucose (TyG) index indirectly assesses insulin resistance, and has been associated with cardiovascular risk. Purpose To investigate whether the TyG index could be a biomarker for assessing the risk of thromboembolic events in patients with AF. Methods We performed an observational prospective cohort study including outpatients diagnosed with AF starting vitamin K antagonist (VKA) or direct-acting oral anticoagulants (DOAC) therapy from January, 2016 to November, 2021. The primary endpoint was a combination of thromboembolic events, considered as the composite of myocardial infarction, and/or venous thromboembolism, and/or ischemic stroke/transient ischemic attack, over 2 years of follow-up. The TyG index was calculated as the natural logarithm (Ln) of the product of baseline plasma glucose and triglyceride using the following formula: Ln [fasting triglycerides (mg/dL) × fasting glucose (mg/dL)]/2. Patients were stratified into tertiles according to their TyG index (ie. first tertile: TyG index <4.59; second tertile: TyG index 4.59-4.83; third tertile: TyG index >4.83). Results We included 2907 patients (median age 77 [70-83] years; 52.5% female). The median TyG index was 4.71 (4.53-4.91), and 34% (989) where in the third tertile of TyG index. During a follow-up of 2 years, 209 (7.2%) patients suffered a thromboembolic event. Compared to patients in the first or second tertiles of TyG index, patients in the third tertile had an increased annual event rate of thromboembolic events (3.40 per 100 patient-years vs. 4.76 per 100 patient-years, p=0.043; and 3.41 per 100 patient-years vs. 4.76 per 100 patient-years, p=0.044), and higher incidence rate ratio (1.41 [95% CI 1.01-1.98], p=0.044 vs. first tertile; 1.40 [95% CI 1.01-1.97], p=0.044 vs. second tertile). Cox regression analyses adjusted for several risk factors demonstrated that the TyG index was an independent predictor for thromboembolic events (aHR 1.82; 95% CI 1.15-2.89; p=0.011). Patients in the third tertile of TyG index presented a 64% higher risk of thromboembolic events (aHR 1.64; 95% CI 1.17-2.29; p=0.004) (Table), with significantly lower event free survival (log-rank test p-value = 0.016) (Figure). Conclusions In this real-world cohort of AF patients under oral anticoagulation therapy, elevated TyG index was an independent risk factor for thromboembolic events. The TyG index is a simple, low-cost indicator, that could be used as a screening tool in everyday clinical practice.