The aim of this study was to examine the frequency, timing, and association of access-site and non–access-site bleeding with mortality in the setting of transcatheter aortic valve replacement (TAVR) during long-term follow-up. Bleeding is frequent and associated with impaired prognosis in patients undergoing TAVR. It is currently unknown whether the site of bleeding differentially influences the outcomes of TAVR patients. In total, 926 consecutive patients undergoing TAVR from 2007 through 2014 were evaluated. Bleeding was assessed according to the Valve Academic Research Consortium 2 criteria. The primary outcome of interest was all-cause mortality up to 5 years of follow-up. A total of 285 patients (30.7%) experienced at least 1 (minor, major, or life-threatening) bleeding event up to 5 years. Compared with patients not experiencing bleeding, the adjusted risk for all-cause mortality was significantly increased among patients with access-site (hazard ratio: 1.34; 95% confidence interval: 1.01 to 1.76; p = 0.04) and non–access-site bleeding (hazard ratio: 2.08; 95% confidence interval: 1.60 to 2.71; p < 0.001). However, non–access-site bleeding conferred a significantly higher risk for mortality compared with access-site bleeding (hazard ratio: 1.56; 95% confidence interval: 1.12 to 2.18; p = 0.009). At multivariate analysis, female sex was a significant correlate of access-site bleeding, whereas chronic kidney disease and the Society of Thoracic Surgeons score were significantly associated with non–access-site bleeding. Among patients with severe aortic stenosis undergoing TAVR, access-site and non–access-site bleeding were independently associated with an increased risk for mortality, with the greatest risk related to non–access-site bleeding during long-term follow-up. Display omitted