Aims: The aim of this study was to compare clinical outcomes of patients undergoing transfemoral transcatheter aortic valve implantation (TAVI) with balloon-expandable (BE) valves vs. self-expandable (SE) valves. Transcatheter aortic valve implantation is a minimally invasive and lifesaving treatment in patients with aortic valve stenosis. Even though BE-valves and SE-valves are both commonly used on a large scale, adequately sized trials comparing clinical outcomes in patients with severe aortic valve stenosis treated with BE-valves compared with SE-valves are lacking. Methods and results: In this CENTER-collaboration, data from 10 registries or clinical trials, selected through a systematic search, were pooled and analysed. Propensity score methodology was used to reduce treatment selection bias and potential confounding. The primary endpoints were mortality and stroke at 30 days follow-up in patients treated with BEvalves compared with SE-valves. Secondary endpoints included clinical outcomes, e.g. bleeding during hospital admission. All outcomes were split for early-generation BE-valves compared with early-generation SE-valves and new-generation BE-valves with new-generation SE-valves. The overall patient population (N= 12 381) included 6239 patients undergoing TAVI with BE-valves and 6142 patients with SE-valves. The propensity matched population had a mean age of 81 ± 7 years and a median STS-PROM score or 6.5% [interquartile range (IQR) 4.0-13.0%]. At 30-day follow-up, the mortality rate was not statistically different in patients undergoing TAVI with BE-valves compared with SE-valves [BE: 5.3% vs. SE: 6.2%, relative risk (RR) 0.9; 95% confidence interval (CI) 0.7-1.0, P = 0.10]. Stroke occurred less frequently in patients treated with BE-valves (BE: 1.9% vs. SE: 2.6%, RR 0.7; 95% CI 0.5-1.0, P = 0.03). Also, patients treated with BE-valves had a three-fold lower risk of requiring pacemaker implantation (BE: 7.8% vs. SE: 20.3%, RR 0.4; 95% CI 0.3-0.4, P < 0.001). In contrast, patients treated with new-generation BE-valves more frequently experienced major and life-threatening bleedings compared with new-generation SEvalves (BE: 4.8% vs. SE: 2.1%, RR 2.3; 95% CI 1.6-3.3, P < 0.001). Conclusion: In this study, which is the largest study to compare valve types in TAVI, we demonstrated that the incidence of stroke and pacemaker implantation was lower in patients undergoing transfemoral TAVI with BE-valves compared with SE-valves. In contrast, patients treated with new-generation BE-valves more often suffered from major or life-threatening bleedings than patients with new-generation SE-valves. Mortality at 30-days was not statistically different in patients treated with BE-valves compared with SE-valves. This study was a propensity-matched analysis generated from observational data, accordingly current outcomes will have to be confirmed in a large scale randomized controlled trial.