Association between Preoperative Vascular Function and Postoperative Arteriovenous Fistula Development

Abstract

Arteriovenous fistula (AVF) maturation failure is the primary cause of dialysis vascular access dysfunction. To evaluate whether preoperative vascular functional properties predict postoperative AVF measurements, patients enrolled in the Hemodialysis Fistula Maturation Study underwent up to five preoperative vascular function tests (VFTs): flow-mediated dilation (FMD), nitroglycerin-mediated dilation (NMD), carotid-femoral pulse wave velocity, carotid-radial pulse wave velocity, and venous occlusion plethysmography. We used mixed effects multiple regression analyses to relate each preoperative VFT to ultrasound measurements of AVF blood flow rate and venous diameter at 1 day, 2 weeks, and 6 weeks after AVF placement. After controlling for AVF location, preoperative ultrasound measurements, and demographic factors (age, sex, race, and dialysis status), greater NMD associated with greater 6-week AVF blood flow rate and AVF diameter (per absolute 10% difference in NMD: change in blood flow rate =14.0%; 95% confidence interval [95% CI], 3.7% to 25.3%; P<0.01; change in diameter =0.45 mm; 95% CI, 0.25 to 0.65 mm; P<0.001). Greater FMD also associated with greater increases in 6-week AVF blood flow rate and AVF diameter (per absolute 10% difference in FMD: change in blood flow rate =11.6%; 95% CI, 0.6% to 23.9%; P=0.04; change in diameter =0.31 mm; 95% CI, 0.05 to 0.57 mm; P=0.02). None of the remaining VFT parameters exhibited consistent statistically significant relationships with both postoperative AVF blood flow rate and diameter. In conclusion, preoperative NMD and FMD positively associated with changes in 6-week AVF blood flow rate and diameter, suggesting that native functional arterial properties affect AVF development.

Keywords: arteriovenous access; arteriovenous fistula; vascular access.

Figure 1.

Postoperative AVF diameter and blood flow are associated with the preoperative FMD and NMD. Ratios of mean AVF blood flow rates and differences in mean AVF vein diameters (millimeters) between the 85th and 15th percentiles of preoperative VFT results using mixed effects regression models. Results are adjusted for AVF location, preoperative ultrasound measures (inflow artery diameter, minimum vein diameter, and brachial artery blood flow), and baseline demographics (age, sex, race, and dialysis status). Expressing the results as comparisons between the 85th and 15th percentiles for each VFT factor allows the strength of the displayed relationships to be compared between the different VFT factors. Results are displayed for day 1 (top line; green), week 2 (middle line; orange), and week 6 (bottom line; brown) ultrasounds. The blue squares represent the estimated ratios of mean blood flow rates or mean differences in diameters for VFTs at the 85th and 15th percentiles, and the colored lines are the 95% CIs. Multiple imputation was performed to impute missing vascular function and ultrasound measurements.

Figure 2.

A greater preoperative NMD is associated with a greater 6-week postoperative AVF blood flow and diameter. (A) The y coordinates on the solid curve provide the ratio of the adjusted geometric mean 6-week AVF blood flow rate at the indicated brachial NMD level to the adjusted geometric mean blood flow rate at the median brachial NMD, which is used as the reference. (B) The y coordinates provide the difference between the adjusted mean 6-week AVF vein diameter at the indicated brachial NMD and the adjusted mean AVF vein diameter at the median brachial NMD reference. Analyses are adjusted for AVF location, baseline ultrasound measurements, and baseline case mix covariates. Multiple imputation was performed to impute missing vascular function and ultrasound measurements. The shaded regions indicate pointwise 95% CIs. Histograms below the curves show the distribution of NMD values in the study cohort.