Risk Prediction of Arteriovenous Fistula Dysfunction in Hemodialysis Patients Using Routine Clinical Indicators

Area of Science:

• Nephrology
• Cardiology
• Clinical Prediction Modeling

Background:

• Arteriovenous fistula (AVF) is the preferred vascular access for hemodialysis (HD).
• AVF dysfunction is a common complication, impacting treatment efficacy.
• Risk factors for AVF patency are not fully understood.

Purpose of the Study:

• To identify key clinical predictors of AVF dysfunction in HD patients.
• To develop a practical model for predicting AVF dysfunction.
• To improve early risk stratification and guide interventions for AVF patency.

Main Methods:

• Retrospective review of 439 HD patients' medical records.
• Utilized Least Absolute Shrinkage and Selection Operator (LASSO) Cox regression for predictor selection.
• Constructed a multivariate Cox proportional hazards regression model for prediction.
• Assessed model discrimination using the concordance index (C-index) and internal validation via bootstrap resampling.

Main Results:

• AVF dysfunction occurred in 10.5% of patients over a median follow-up of 2.9 years.
• Five significant predictors identified: total protein, albumin, left ventricular ejection fraction (LVEF), history of hypertension, and history of heart disease.
• The final prediction model demonstrated strong discrimination with a C-index of 0.812 (95% CI: 0.753-0.871).

Conclusions:

• Established five routinely available clinical variables as independent predictors of AVF dysfunction.
• Developed a nomogram with high predictive accuracy for AVF dysfunction.
• The model can aid in early risk stratification and timely interventions to prevent AVF failure and enhance dialysis efficacy.