Development of a prediction model to estimate short- and long-term clinical outcomes in arteriovenous grafts for hemodialysis patients.

BACKGROUND: The prognosis and quality of life for patients undergoing hemodialysis (HD) are significantly compromised by dysfunctions in arteriovenous grafts (AVG). This study aimed to develop and validate a risk prediction model for factors influencing both short- and long-term primary patency rates in HD patients with AVG. METHODS: The single-center retrospective study comprised 354 HD patients who had undergone AVG creation. Using Kaplan-Meier survival analysis and Cox regression models, we evaluated the patency rates of AVG and identified associated independent risk factors based on retrospectively collected clinical data. RESULTS: The primary patency rates of the patients with AVG at 6 months, 1 year, and 2 years were 82.49%, 50%, and 16.95%, respectively. The multivariate Cox regression analysis revealed that the brachial arterial blood flow and draining venous diameter at 1 month post-operatively were independent predictors of newly created AVG patency at 6 months. Furthermore, at 1 month post-operatively, the brachial artery blood flow, draining vein diameter, and history of central venous catheter (HR 0.480 (0.281, 0.820), p = 0.007) were independent risk predictors for failure of newly created AVG at 12 months. CONCLUSION: This study revealed that these parameters, including brachial arterial blood flow, draining vein diameter, and history of central venous catheter, could be applied to predict short- and long-term risks for AVG patency loss.