Agreement Between Transthoracic Echocardiography and Digital Subtraction Angiography for the Detection of Right Atrial Thrombus in Patients Undergoing Maintenance Hemodialysis: A Retrospective Cohort Study

Background/Objectives: Patients on maintenance hemodialysis (MHD) have an increased risk of catheter-related right atrial thrombosis (CRAT) due to long-term central venous catheterization. Transthoracic echocardiography (TTE) and digital subtraction angiography (DSA) are commonly used to evaluate right atrial thrombi, yet their diagnostic consistency in this patient population remains unclear. This study aimed to compare TTE and DSA diagnostic agreement for CRAT and explore its independent risk factors. Methods: We retrospectively enrolled 254 MHD patients who underwent paired same-day TTE and DSA from November 2021 to May 2025. Cohen’s kappa, intra-class correlation coefficient (ICC), and Bland–Altman analysis were used to evaluate diagnostic consistency. Univariable and multivariable logistic regression screened independent CRAT risk factors. Results: The thrombus detection rates of TTE and DSA were 17.3% and 14.6%, respectively, with moderate diagnostic consistency (κ = 0.487, p < 0.001). The ICC values were 0.871 for thrombus length and 0.824 for thrombus width between TTE and DSA, indicating excellent consistency in thrombus size measurement between the imaging modalities. Catheter tip located in the mid-right atrium, decreased central venous catheter flow, prolonged catheterization, and male sex were independent risk factors for CRAT. The predictive model exhibited good discrimination (area under the curve = 0.788) and acceptable calibration (Hosmer–Lemeshow, p = 0.429). The optimal cutoff derived from the Youden index was 0.191, with a corresponding sensitivity of 81.0% and specificity of 67.3%. Conclusions: TTE is recommended as the first-line CRAT screening tool for MHD patients. Combined TTE and DSA use improves diagnostic yield. Regular targeted surveillance for high-risk individuals reduces CRAT incidence.