Background Clinically significant steal syndrome occurs in a subset of dialysis patients with arteriovenous (AV) access. Factors associated with steal are poorly understood. Severe symptoms require access revision or sacrifice, potentially jeopardizing access options. Our objective was to review our dialysis access experience to identify factors associated with significant steal syndrome. Methods We reviewed all adult patients undergoing their first permanent upper extremity access, AV fistula (AVF) or AV graft (AVG), between January 2008 and July 2011 at a single center. Medical, demographic, and access characteristics were collected from our electronic medical record and a local dialysis center's database. Patients who required correction of steal syndrome were compared with the larger access cohort. Statistical analysis included Fisher's exact test and χ2 for noncontinuous variables and unpaired t-test for continuous variables. Results Of the 303 patients, 15 required correction for steal syndrome (8 of 232 AVF and 7 of 71 AVG). Eight were ligated; 2 were initially banded, then ligated; and 5 underwent distal revascularization with interval ligation. Coronary artery disease was more prevalent in steal syndrome patients (66.7% vs. 25%, P = 0.001); the same was found with peripheral arterial disease (40% vs. 13.8%, P = 0.02). Furthermore, more patients with steal syndrome were on clopidogrel for cardiovascular reasons (40% vs. 9%, P = 0.002). Steal syndrome only developed with AVF and AVG using brachial artery inflow. No cases of steal syndrome arose from radial/ulnar inflow (P = 0.03). All AVG with steal syndrome had a straight configuration; no looped AVG developed steal (P = 0.02). Other patient characteristics such as age, sex, race, hypertension, diabetes mellitus, congestive heart failure, cerebrovascular accident, cause of end-stage renal disease, and other medication history were not different between groups. Conclusions Clinically significant steal syndrome is associated with disease in coronary and peripheral arterial beds. In addition, the use of brachial artery inflow and straight AVG configuration is associated with steal syndrome. Consideration should be given to construction of access using smaller forearm arteries and looped AVG configuration in patients with high risk for steal. In addition, such patients may require more vigilant monitoring for development of steal after access construction.