Abstract: 【Abstract】 Objective Method We recruited 1,051 end-stage renal disease patients in our hospital for this study. Their main factors and regression coefficients relating to AVF surgery were collected, to establish a risk score model based on the regression coefficient, to create a prediction score formula, and to evaluate their efficiencies. Results＜0.0001). It appears that the prediction score of 103.17 may be an acceptable cut off value, yielding a sensitivity of 69.46%, specificity of 68.42%, positive predictive value of 98%, and negative predictive value of 9% in a subject population.  Conclusions The prediction score formula, established based on cephalic vein diameter, radial artery diameter, serum total cholesterol, the average diastolic blood pressure 2 days prior to the surgery to the morning of the surgery day, may be useful for the prediction of AVF early patency rate. Logistic regression analyses demonstrate that cephalic vein diameter (t=2.202, OR=2.085), radial artery diameter (t=3.493, OR=1.767), serum total cholesterol (t=-2.281, OR=0.708), and the average diastolic blood pressure 2 days prior to the surgery to the morning of the surgery day (t=2.723, OR=1.057) are the major factors relating to the success rate of AVF. The area under ROC curve based on individual total prediction scores calculated from risk score model is 0.707 (P To establish a simple prediction model for the early patency rate of arteriovenous fistula (AVF).

Key words: Arteriovenous fistula, Prediction of success rate, Risk score model, Prediction score formula