Preoperative Vascular Medial Fibrosis and Arteriovenous Fistula Development

Abstract

Background and objectives: Arteriovenous fistula maturation requires an increase in the diameter and blood flow of the feeding artery and the draining vein after its creation. The structural properties of the native vessels may affect the magnitude of these changes. We hypothesized that an increase in the collagen content of the vascular media (medial fibrosis) preoperatively would impair vascular dilation and thereby, limit the postoperative increase in arteriovenous fistula diameter and blood flow and clinical arteriovenous fistula maturation.

Design, setting, participants, & measurements: We enrolled 125 patients undergoing arteriovenous fistula creation between October of 2008 and April of 2012 and followed them prospectively. Any consenting subject was eligible. Arterial and venous specimens were sampled during arteriovenous fistula surgery. Masson's trichrome-stained samples were used to quantify medial fibrosis. Arteriovenous fistula diameter and blood flow were quantified using 6-week postoperative ultrasound. Clinical arteriovenous fistula maturation was assessed using a predefined protocol. The association of preexisting vascular medial fibrosis with arteriovenous fistula outcomes was evaluated after controlling for baseline demographics, comorbidities, and the preoperative venous diameter.

Results: The mean medial fibrosis was 69%±14% in the arteries and 63%±12% in the veins. Arterial medial fibrosis was associated with greater increases in arteriovenous fistula diameter (Δdiameter =0.58 mm; 95% confidence interval [95% CI], 0.27 to 0.89 mm; P<0.001) and arteriovenous fistula blood flow (Δblood flow =85 ml/min; 95% CI, 19 to 150 ml/min; P=0.01) and a lower risk of clinical arteriovenous fistula nonmaturation (odds ratio, 0.71; 95% CI, 0.51 to 0.99; P=0.04), all per 10% absolute difference in medial fibrosis. In contrast, venous medial fibrosis was not associated with the postoperative arteriovenous fistula diameter, blood flow, or clinical maturation.

Conclusions: Preoperative arterial medial fibrosis was associated with greater arteriovenous fistula diameter and blood flow and a lower risk of clinical arteriovenous fistula nonmaturation. This unexpected observation suggests that medial fibrosis promotes arteriovenous fistula development by yet undefined mechanisms or alternatively, that a third factor promotes both medial fibrosis and arteriovenous fistula maturation.

Keywords: Arteriovenous Shunt, Surgical; Collagen; Demography; Humans; Odds Ratio; Tunica Media; Veins; arteries; arteriovenous fistula; fibrosis; hemodialysis.

Figure 1.

Distribution of medial fibrosis. Distribution among samples of the (A) arteries and (B) veins used to create arteriovenous fistulas.

Figure 2.

Collagen and elastin in samples of the arteries used to create arteriovenous fistulas. Masson’s trichrome stain shows (A) an artery with mild medial fibrosis (49%) and (B) an artery with significant medial fibrosis (95%). (C and D) Elastic van Gieson stain of these two samples shows that both arteries have little elastin in their respective medial layers. Collagen appears blue and smooth muscle cells appear red in Masson’s trichrome stain. Elastic lamellae appear black in elastic van Gieson stain.

Figure 3.

Representative second harmonic generation images of the different patterns of medial collagen fiber orientation. (A) Parallel to lumen, (B) Perpendicular to lumen, (C) Random, (D) Railroad track, (E) Honeycomb, and (F) Mixed patterns. The relative frequency of each pattern is summarized in Table 2. The microperpendicular with macroparallel pattern, which accounted for very few samples, is not displayed.

Figure 4.

Association of arterial medial fibrosis with 6-week arteriovenous fistula (AVF) diameter and blood flow. (A) The y coordinate of the red curve provides the difference between the adjusted mean 6-week AVF diameter at the indicated arterial fibrosis and the adjusted mean 6-week AVF diameter at the median arterial fibrosis, which is used as the reference. (B) The y coordinate provides the ratio of the adjusted geometric mean 6-week AVF blood flow rate at the indicated arterial fibrosis to the adjusted geometric mean 6-week AVF blood flow rate at the median arterial fibrosis reference. Analyses were adjusted for preoperative vein diameter and baseline case mix covariates. Multiple imputation was performed to impute missing arterial fibrosis and ultrasound measurements. The shaded regions indicate pointwise 95% confidence intervals.

Figure 5.

No association between arterial medial fibrosis and the arterial elastic modulus (a measure of arterial stiffness). The regression coefficient for a linear regression is −0.14 (95% confidence interval, −0.58 to 0.30) per 10% absolute difference in medial fibrosis (P=0.52).