Construction of risk factors and prediction model for arteriovenous graft thrombosis

Objective: This study aims to identify risk factors for thrombosis in arteriovenous grafts and construct a predictive model to assess thrombosis risk in patients undergoing maintenance hemodialysis (MHD). Methods: A total of 160 MHD patients with arteriovenous graft were included and divided into a thrombosis group (n=39) and a control group (n=121). Univariate and multivariate logistic regression analyses were performed to identify independent risk factors. A nomogram prediction model was developed using R software, and its predictive performance was evaluated through calibration curves and C-index validation. Results: Multivariate analysis identified diabetes, hypotension during dialysis, arteriovenous graft stenosis, compression hemostasis >30 min, and calcium-phosphorus product >55 mg²/dL² as independent risk factors for arteriovenous graft thrombosis. The nomogram model demonstrated good predictive accuracy, with an initial C-index of 0.753 and a validated C-index of 0.735. Conclusion: The established nomogram effectively predicts arteriovenous graft thrombosis risk, aiding early identification and targeted intervention for high-risk patients.