OBJECTIVE: Complications from vascular access account for 15% of hospital admissions among US hemodialysis patients. Complications are less frequent with arteriovenous fistulas than with synthetic grafts. We assessed clinical and nonclinical predictors of whether patients with end-stage renal disease (ESRD) starting hemodialysis receive a fistula or graft. We also investigated changes in practice between 1986-1987 and 1990. DESIGN: Cross-sectional study. SETTING: United States hemodialysis population. PATIENTS: Random, national samples of ESRD patients who started hemodialysis in 1986-1987 (n=2741) or 1990 (n=1409) from United States Renal Data System Special Studies. MAIN OUTCOME MEASURE: Type of permanent vascular access (arteriovenous fistula vs synthetic graft), analyzed using multivariate logistic regression. RESULTS: Clinical and demographic factors as well as socioeconomic status, region of residence, and year starting hemodialysis predicted the type of vascular access. Overall, 56% of patients had grafts 30 days after starting dialysis, but graft use increased from 51% in 1986-1987 to 65% in 1990 (adjusted odds ratio [AOR], 1.67 for 1990 vs 1986-1987; 95% confidence interval [CI], 1.43-1.95; P<.001). Graft use (relative to fistula) varied by region of residence (ranging from AOR, 0.20; 95% CI, 0.14-0.28; P<.001 [New England], to AOR, 2.69; 95% CI, 2.03-3.58; P<.001 [East South Central]; both relative to the national average). CONCLUSIONS: This national study documents large variations in the relative use of fistulas and grafts and a trend away from fistulas. The prevalence of comorbid conditions fails to explain these findings. Presentation and referral of patients early in the process of their ESRD, teaching surgeons to place fistulas, and training dialysis nurses to access fistulas may increase their use.
OBJECTIVE: Complications from vascular access account for 15% of hospital admissions among US hemodialysis patients. Complications are less frequent with arteriovenous fistulas than with synthetic grafts. We assessed clinical and nonclinical predictors of whether patients with end-stage renal disease (ESRD) starting hemodialysis receive a fistula or graft. We also investigated changes in practice between 1986-1987 and 1990. DESIGN: Cross-sectional study. SETTING: United States hemodialysis population. PATIENTS: Random, national samples of ESRDpatients who started hemodialysis in 1986-1987 (n=2741) or 1990 (n=1409) from United States Renal Data System Special Studies. MAIN OUTCOME MEASURE: Type of permanent vascular access (arteriovenous fistula vs synthetic graft), analyzed using multivariate logistic regression. RESULTS: Clinical and demographic factors as well as socioeconomic status, region of residence, and year starting hemodialysis predicted the type of vascular access. Overall, 56% of patients had grafts 30 days after starting dialysis, but graft use increased from 51% in 1986-1987 to 65% in 1990 (adjusted odds ratio [AOR], 1.67 for 1990 vs 1986-1987; 95% confidence interval [CI], 1.43-1.95; P<.001). Graft use (relative to fistula) varied by region of residence (ranging from AOR, 0.20; 95% CI, 0.14-0.28; P<.001 [New England], to AOR, 2.69; 95% CI, 2.03-3.58; P<.001 [East South Central]; both relative to the national average). CONCLUSIONS: This national study documents large variations in the relative use of fistulas and grafts and a trend away from fistulas. The prevalence of comorbid conditions fails to explain these findings. Presentation and referral of patients early in the process of their ESRD, teaching surgeons to place fistulas, and training dialysis nurses to access fistulas may increase their use.