Objective: A total 428 HD patients were prospectively recruited in K-cohort from June 2016 to March 2020. Mean follow-up period was 28 months. The serum endostatin, follistatin1 (FST1), galectin-3, and matrix metalloproteinase-2 (MMP-2) at the time of study enrollment were measured using enzyme-linked immunosorbent assay (ELISA).  Methods: Multivariable cox regression analysis showed higher log endostatin was significantly associated with cardiac mortality. The cut-off value of endostatin for prediction of cardiac death was determined as 161.2ng/mL using univariable cox regression analysis. We divided the patients as a high (n=244) and low (n=184) endostatin group according to plasma endostatin level of 161.2ng/mL. The high endostatin group presented longer HD duration, more ultrafiltration volume, higher predialysis systolic blood pressure and, higher β2-microglobulin than low endostatin group. In addition, cardiovascular ischemia or inflammatory markers, such as FST1, galectin 3 and MMP2, were significantly higher in high endostatin group. Cox regression analysis revealed that cardiac event (HR=6.51, p=0.01), cardiac death (HR 6.59, p=0.01), and all cause death (HR 2.32, p=0.01) was markedly increased in high endostatin group after adjusting age, BMI, diabetes, hypertension, LDL cholesterol, and use of statin. Results: The HD patients with high circulating endostatin (more than 161.2ng/mL) are likely to occur cardiac event, and cardiac death. Higher plasma endostatin might be a sensitive marker to predict cardiac outcomes Conclusions: Objective: Endostatin is a C-terminal fragment of type XVIII collagen released during extra cellular matrix remodeling. It was suggested as a biomarker to predict cardiovascular morbidity and mortality. We aim to clarify that circulating endostatin can predict cardiac mortality in hemodialysis (HD) patients.  Methods: A total 428 HD patients were prospectively recruited in K-cohort from June 2016 to March 2020. Mean follow-up period was 28 months. The serum endostatin, follistatin1 (FST1), galectin-3, and matrix metalloproteinase-2 (MMP-2) at the time of study enrollment were measured using enzyme-linked immunosorbent assay (ELISA).  Results: Multivariable cox regression analysis showed higher log endostatin was significantly associated with cardiac mortality. The cut-off value of endostatin for prediction of cardiac death was determined as 161.2ng/mL using univariable cox regression analysis. We divided the patients as a high (n=244) and low (n=184) endostatin group according to plasma endostatin level of 161.2ng/mL. The high endostatin group presented longer HD duration, more ultrafiltration volume, higher predialysis systolic blood pressure and, higher β2-microglobulin than low endostatin group. In addition, cardiovascular ischemia or inflammatory markers, such as FST1, galectin 3 and MMP2, were significantly higher in high endostatin group. Cox regression analysis revealed that cardiac event (HR=6.51, p=0.01), cardiac death (HR 6.59, p=0.01), and all cause death (HR 2.32, p=0.01) was markedly increased in high endostatin group after adjusting age, BMI, diabetes, hypertension, LDL cholesterol, and use of statin. Conclusions: The HD patients with high circulating endostatin (more than 161.2ng/mL) are likely to occur cardiac event, and cardiac death. Higher plasma endostatin might be a sensitive marker to predict cardiac outcomes