OBJECTIVESArteriovenous fistula (AVF) is the preferred type of vascular access for hemodialysis to treat end-stage renal disease. A high proportion of AVF are never used for dialysis because the vein fails to mature adequately. We have previously described the safety and feasibility of Vascugel (Genzyme BioSurgery, Cambridge, Mass) (allogeneic aortic endothelial cells in a gelatin matrix) when placed around the anastomotic and venous outflow sites of AVFs (Vascular intimal Hyperplasia: Extending Arterial and venous patency, Limiting vascular Trauma, and inhibiting Hyperplasia while re-establishing vascular health [V-HEALTH] clinical study). In this retrospective analysis, we investigated factors that influenced AVF remodeling in patients from the V-HEALTH study. We hypothesized that providing healthy endothelial cells and their secreted factors immediately after surgery could enhance venous remodeling in the setting of vascular injury.METHODSThirty-one AVF patients from the V-HEALTH study were randomized 2:1 to receive either Vascugel or control matrices (placebo) at surgery and were followed for 24 weeks. Venous lumen diameter was measured by ultrasound at 1, 3, and 5 cm from the anastomosis. Vein remodeling (change in lumen diameter at 4, 12, and 24 weeks compared with baseline diameter at 2 weeks) was analyzed using a multiple regression mixed model.RESULTSThe results indicated that diabetes was a significant, negative predictor of venous remodeling over the 24-week study (P = .02). The model-predicted change in lumen diameter from 2 to 24 weeks was -0.7 mm in diabetic patients (n = 11) and +2.4 mm in nondiabetic patients (n = 15), a difference of 3.1 mm, 95% confidence interval [CI] (1.4-4.9), P = .0014. Patient race, baseline vein diameter, and time post-AVF creation were also significant factors that affected remodeling (P < .05). Compared with placebo, there was a strong suggestion that Vascugel treatment improved the rate of venous enlargement in diabetic patients (P = .05). The model-predicted change in lumen diameter at 24 weeks was -1.9 mm for placebo-treated diabetic patients and +0.4 mm for Vascugel-treated diabetic patients, a difference of 2.3 mm, 95% CI (-0.1-4.8), P = .06, suggesting that treatment with Vascugel may mitigate the negative influence of diabetes on AVF remodeling.CONCLUSIONSDiabetes negatively impacts AVF remodeling and targeted local therapy with perivascular, allogeneic endothelial cells may ameliorate this effect. A phase II trial designed specifically to evaluate AVF remodeling is needed to determine if Vascugel can increase AVF maturation and use and to support larger randomized trials.