Guidelines recommend that > or =50% of patients starting dialysis have a fistula. We reviewed our experience in consecutive incident patients over a 1-year period. Only 30 of the 93 patients starting hemodialysis had a fistula that was accessed. Late referral (nephrology contact <90 days) was a significant issue in 48% (30/63) of the patients without a fistula. Most (n=21) of the late referrals were acute disease; only 9 were late referrals of chronic disease. Nephrology follow-up exceeded 200 days in the remaining (33/63) without this access. In the cohort with sufficient nephrology referral, we explored variables associated with a fistula (n=30) compared with those without one (n=33). In multivariate logistic regression analysis, peripheral vascular disease (odds ratio [OR] 0.026, 95% confidence interval [CI] 0.002-0.286) and rapid loss of estimated glomerular filtration rate (eGFR) (OR 0.745 per mL/min/1.73 m(2)/year, 95% CI 0.625-0.888) in the year preceding dialysis were significant negative predictors for a fistula. Patients without access experienced faster declines in GFR in the year preceding dialysis (12.1+/-9.9 vs. 4.7+/-3.5 mL/min 1.73 m(2) with access, p<0.001). Glomerular filtration rate loss in the 2 years before starting dialysis was the same between the 2 groups (-0.54+/-10.4 vs. 1.42+/-3.9 mL/min 1.73 m(2)). Age, sex, diabetes, other comorbidity, length of nephrology follow-up, eGFR at dialysis start, hemoglobin, and albumin were not significant. At our center, rapid loss of renal function in otherwise stable chronic kidney disease (CKD) patients is more important than late referral of CKD for the lack of access. Improvements in rapid referral for access creation could help reduce this barrier.