Background. In the United States, the use of polytertraflouroethylene (PTFE) graft compared with native arteriovenous fistula (AVF) for haemodialysis vascular access has been increasing despite a greater than twofold higher incidence of thrombosis and infection associated with PTFE grafts. Methods. We studied 214 haemodialysis patients with not more than two revisions of their vascular access, to determine whether any relationship exists between the type of haemodialysis vascular access and dialysis dose assessed primarily by urea reduction ratio (per cent reduction in blood urea nitrogen concentration after a dialysis session). Serum albumin concentration was used as a secondary outcome measure of dialysis adequacy. Urea reduction ratio and predialysis serum albumin concentration were measured at onset of study and at 4-week intervals and mean values were calculated for each subject. Results. The 214 patients (118 males, 96 females) included 173 Blacks (81%), 26 Whites (15%), and 15 Hispanics (7%), of mean (+/- SD) age 55.6 +/- 15.5 years. Of these 214 subjects, 111 (52%) had a native AVF, while 103 (48%) had a PTFE graft. Both mean urea reduction ratio (native AVF=69+/-6.7% vs PTFE graft = 70 +/- 7.3%; P = 0.31), and mean serum albumin concentration (native AVF = 4.02 +/- 0.39 g/dl vs PTFE graft = 4 +/- 0.33 g/dl; P = 0.59) were equivalent in both groups. Separate multiple logistic regression analyses with type of vascular access as one of the independent variables, found no significant relationship between type of vascular access and either a urea reduction ratio > 65% (P = 0.67), or a serum albumin concentration >4 g/dl (P = 0.89), after adjustment for age of vascular access, access revision, location of access, dialyser urea clearance, length of dialysis treatment, body weight, and age. Conclusion. We conclude that PTFE grafts do not permit delivery of better dialysis than native AVF. The increasing use of PTFE grafts in the United States does not have any clinical justification.